94 Colum. L. Rev. 2594 (1994)


Dennis S. Karjala [FNa]


The digital age brings with it increasing varieties of products that are costly to produce but arguably vulnerable to cheap and rapid copying. An absence of legal protection acts as a disincentive to investment in the development of such products. Since patent law protects only nonobvious advances in applied functionality and traditional copyright law abjures the protection of function, products of human intellectual activity falling outside these two primary regimes of protection have historically been left to their fate under trade secret and unfair competition law or, failing that, to the vagaries of the free market. The question before us is whether a third general paradigm can be constructed for products inadequately or inappropriately protected by patent or copyright law that better balances creation incentives against free use and efficient development.

In the first principle article of this Symposium, Legal Hybrids Between the Patents and Copyright Regimes, [FN1] Professor Reichman proposes a third intellectual property paradigm to cover "legal hybrids." The second principle article, A Manifesto Concerning Legal Protection of Computer Programs, [FN2] applies that theory to computer software. Both articles use the term "market failure" to refer to the underproduction of products resulting from too low a level of protection under intellectual property law. [FN3] These two primary Symposium articles correctly identify "market failure" as the central problem for the protection of products that fall through the crack between patent and copyright. This alone is an important contribution to the ongoing dialogue on how to optimize intellectual property rights in such products for the overall benefit of society.

Unfortunately, neither paper concretely tells us how to solve the problem. Three characteristics of the papers explain this failure. First, *2595 both papers use analytically confusing or inapt catchphrases [FN4] that deflect attention from the market failure notion (or, as I use the term below, "misappropriation") that should form the cornerstone of the analysis. Second, the papers make no attempt to demonstrate how to identify "market failures" or, in the case of the Manifesto, to show that any market failure has actually occurred for software (outside of program code, which is adequately handled by copyright). Third, and most important, even the general features of the third paradigm suggested by both papers are simply variations of the traditional dimensions of patent or copyright (objects protected, term of protection, scope of protection, independent creation, registration, compulsory licensing), without supplying any concrete choices along any of these dimensions. The wide variety of sui generis approaches to which Professor Reichman points [FN5] attests to the inability of a single scheme based solely on variations along these traditional dimensions to work in general.

In contrast to the catchphrases used in the principle articles, I suggest analyzing the problem under another intellectual property term: misappropriation. By "vulnerability to misappropriation" I mean that competitive products can cheaply and rapidly be brought into production after the original appears on the market and before the original creator has a chance to recover investment costs. While "misappropriation" so defined is similar to the concept of "market failure" adopted by Professor Reichman and the authors of the Manifesto, it is analytically more powerful, because it does not focus on specific objects of protection (such as "industrial compilations," "know- how," or "algorithms") but rather looks to methods of acquiring information that, if permitted, would result in disincentives to create desirable works. A focus on misappropriation, for example, would exclude from the new scheme of protection any information-based product that is not the subject of market failure, whether or not it qualifies as an "industrial compilation" or a "legal hybrid," and would include products (such as nonselective databases [FN6]) that are subject to market failure notwithstanding that nonfunctionality disqualifies them as "legal hybrids." [FN7]

*2596 Whether a single general antimisappropriation statute can be written that will cover the "market failures" in need of regulation goes beyond the scope of this Comment. The Comment does, however, suggest the new conceptual direction in which we should be looking for a third paradigm. After much time and effort, neither of the primary papers to this Symposium offers a concrete solution to the problem of protecting any given "legal hybrid," let alone a general solution for all products subject to market failure. Even in the more limited area of computer software, the Manifesto defines no objects of protection, demonstrates no market failures for any "industrial compilations" (beyond code) associated with computer programs, provides no clear statement of the scope of protection, offers no concrete term of protection, and suggests no specifics concerning the problems of injunctions and compulsory licensing. Professor Reichman, too, leaves these details to negotiation among the "relevant technical communit[ies]." [FN8] The failure of these two monumental efforts to find a meaningful solution using such traditional intellectual property notions suggests the need to look in new directions. This Comment proposes that a focus on unfair methods of copying, as opposed to the fact of copying (or the nature of the work copied), is potentially a fruitful place to begin.

Part I of this Comment takes functionality as the dividing line between the traditional universes of patent and copyright subject matter and explains how market failure can result from this traditional division of labor. It also explains how those market failures inhere in unfair methods of copying (as opposed to the nature of the products copied). Part I goes on to consider the possibility of extending copyright to cover future market failures relating to software, without adopting a full-blown new paradigm. Part II outlines the primary issues involved in creating a third intellectual property paradigm based on the concept of misappropriation. It first illustrates how a focus on methods of copying better directs us toward appropriate regulatory goals in cases of market failure than the approaches of the two primary papers. It then addresses more concretely the questions that must be faced in adopting the new paradigm.

                                                                    I. Patent, Copyright, and Misappropriation

A. Market Failure in the Gap Between Patent and Copyright

Market failure can arise because the traditional gap between patent and copyright pointed out by Professor Reichman [FN9] includes some products that are costly to produce initially but may be cheaply and easily copied. This Section will show how the basic difference in regulatory goals between patent and copyright leads at times to market failure. The market*2597 failure occurs not simply because a product is copied by a competitor but rather because copying is effected by cheap and easy methods that allow undercutting the original creator's price before she can recover her investment. If those unfair methods can be regulated, other methods of copying, and therefore freer competition, can still be permitted.
Functionality determines the traditional division of labor between patent and copyright and explains the evolution of these two very different schemes of protection for the products of human creativity. [FN10] That is to say, patent law protects creative but functional invention, while copyright protects creative but nonfunctional authorship. The policy basis for the distinct approaches of patent and copyright law is the social desirability of allowing later technological creators -- creators of functional works -- to build on and to improve, often incrementally, the earlier works of others. [FN11]

These two main paradigms are designed to leave some products of human creativity unprotected. Ideas and theories are not protected by copyright, and incremental technological advances are not protected by patent, notwithstanding that both are often creative. The creator of a historical theory may find another writer using the theory to entertain a broader audience at a handsome profit, and the creator of an unpatentable functional product has a monopoly only for the period that is required for competitors to recognize the value or popularity of the improved product, to figure out its "secret," if any, and to gear up for production and marketing. These deliberate results reflect a fear that greater protection would hinder the development of more and better products and works than it would engender.

*2598 Nevertheless, advances in technology at times upset the carefully drawn balances of patent and copyright, even as supplemented by trade secret law. Patent protection may be unavailable for a certain desirable product because its know-how represents an incremental technological advance. Copyright protection may also be unavailable because the product is functional, factual, or "noncreative." Absent intellectual property protection for the information or the know-how associated with such products, there may be a disincentive to produce them, because the cost of copying the information or know-how using new technology is below the cost of creating the original product.

B. Unfair Methods of Copying as Market Failure

One legislative approach to the problems posed by such market failures has been a haphazard series of sui generis schemes, as outlined by Professor Reichman, each focusing on specific objects of protection such as utility models or plant varieties. [FN12] If we follow Professor Reichman and take as given that these sui generis schemes have not worked well, the question is whether another approach will work better. This Comment suggests focusing on unfair methods of copying. There are in fact at least some examples in which courts or legislatures have tried to focus, as such, on unfair methods of taking or redelivering otherwise unprotected information. The original misappropriation case [FN13] arose because the new technology of telegraphy permitted fast and inexpensive copying across the continent of hard-won information that was freely copyable under the primary paradigms. A Florida statute outlawing plug-mold copying of boat hulls [FN14] was an effort to regulate only one means of taking another's otherwise unprotected industrial design, a particularly misappropriative means that was the three-dimensional equivalent of photocopying. Even the Manifesto 's notion of "trivial acquisition of behavioral equivalence" focuses by its terms on unfair methods of such acquisition, rather than on acquisition of behavioral equivalence per se. [FN15] This Comment tries to look more generally at how we can place some statutory order on unfair methods of copying without the limitations associated with sui generis legislation focused on specific types of products or pure common-law judicial developments.

*2599 Examination of a complete (nonselective) electronic database containing public domain elements, such as LEXIS, provides a good example of the need to focus on methods of copying as opposed to the nature of the work copied. While such databases are not "legal hybrids" under Professor Reichman's analysis, they are now vulnerable to misappropriation as a result of the rejection of the "sweat of the brow" basis for copyright originality. [FN16] Their vulnerability, however, comes from the ease of electronic copying, rather than copying by means of reinputting the public domain contents case-by-case into a new database by hand. If a competitor were to call each case out of LEXIS to a computer screen and re-enter it by hand, copying from the screen, he would have to expend essentially the same effort in labor and proofreading as LEXIS did in creating the database in the first place. The competitor would then have little or no cost savings with which to undercut LEXIS's efforts to get a return on its initial investment. Because the cases themselves are in the public domain, there is no reason in this situation for the law to regulate the competitor's copying activity (that is to say, there is no "market failure"). If, however, the competitor could download the LEXIS database, even one case at a time, the savings would be enormous. Consequently, we cannot expect copyright-unprotected database creators to make them available in a form in which such electronic copying is possible. A similar market failure occurs for digitally formatted versions of public domain texts, such as the complete works of Shakespeare. Unless we can protect against cheap and easy methods of copying, we are left with the unpalatable choice of allowing misappropriation and the concomitant underproduction (or "underavailability") of desirable works or of recognizing a new copyright in the underlying works, essentially removing them from the public domain. [FN17]

C. Extending Copyright to Protect Against Misappropriation of Software

New intellectual property legislation is always the subject of fierce debate even at the domestic level. It is therefore very difficult to overcome special interest lobbying and achieve a result that optimally balances the overall public interest. Moreover, when the legislation represents *2600 a new paradigm, achieving consensus at the international level can take a very long time. Because this Symposium focuses on software, this Section undertakes a brief consideration of the extent to which copyright can be interpreted to deal adequately with software market failures. Part II will then consider the issues involved in creating a general antimisappropriation statute that would achieve the regulatory goal more broadly.

In earlier studies I have used the misappropriation concept -- what the primary papers here term "market failure" -- to determine the appropriate interpretation of copyright law as applied to computer software. [FN18] Indeed, it is precisely the vulnerability of program code to misappropriation that justifies a copyright-like protection scheme. Provided the scope of protection is appropriately narrow, the use of copyright to protect such a functional work seems to do little harm. The question is whether copyright can be extended to cover other market failures associated with software, assuming the existence of such market failures is demonstrated.

It would be a mistake to protect, as such, functional aspects of software other than program code under the minimal creativity threshold and long protection period of copyright. General copyright protection of program designs or specific collections of program functions, for example, could inhibit the asymptotic advance toward optimally efficient solutions that comes from the incremental contributions of many (often anonymous) systems engineers and programmers, as well as feedback from users. However, if technological development eventually makes noncode aspects of software vulnerable to piracy, a limited application of the misappropriation principle to the protection of program code might solve the problem without a sui generis statute. For example, in CMAX/Cleveland, Inc. v. UCR, Inc., [FN19] the defendants used a code generator to create their own program code directly from the screen images produced by plaintiff 's program. Because the screen images themselves were the product of plaintiff 's code, it would be reasonable to interpret defendant's particular method of programming as resulting in an indirect, and therefore infringing, copy of plaintiff 's code. [FN20]

If the courts were expressly to recognize copyright as an antimisappropriation safeguard for software, as opposed to a device solely for the protection of literary or artistic creativity, many if not all of the future code-related market failure problems relating to software might be solved, in the manner of the CMAX case. Copyright protection of software would be considered "thin," and the copyright would be infringed only by misappropriative methods of copying, that is, methods of cheap and easy code copying whose widespread use would result in disincentives *2601 to creation. Nevertheless, dealing with future market failures involving program structure or user interfaces under copyright runs a serious risk of also protecting functional elements that are not subject to market failure. [FN21] Moreover, even if copyright could meaningfully be interpreted to cover the market failures occurring for computer software, statutory amendment would be required for such works as nonselective databases, [FN22] boat hulls and other functional designs, [FN23] and semiconductor chips to the extent that the protection of mask works is superseded by advanced technology. [FN24] Consequently, a new paradigm more explicitly focused on the market failure problem is worth considering.

                                                          II. Issues in Creating a General Antimisappropriation Statute

This Part first shows that the goal of regulating only those methods of copying that lead to market failure is hardly advanced, and may even be inhibited, by reliance on some of the catchphrases used by the primary *2602 contributions to this Symposium. It then goes on more positively to look at the substantive issues involved in developing a statutory third paradigm based on the misappropriation/market failure concept.

A. Maintaining the Focus on Market Failure

The feasibility of a general antimisappropriation statute depends on an operational definition of "misappropriation" that captures all or at least a significant number of true market failures. At the same time, the definition must maintain the general presumption of the major paradigms against protection of unpatentable functionality and copyright-unprotected factual information. Some of the concrete issues involved in developing a concrete operational definition of misappropriation are taken up in the next Section. Here I note that, in seeking such a definition, we should stay focused on market failure by avoiding the catchphrases that permeate both of the primary papers to this Symposium. The key is not that products need protection to the extent they "bear their know-how on their face" or "near the surface." Visual inspection may reveal the functional value of the shape of an unpatented industrial product, for example, but the product normally will not be easily duplicated by just anyone. Even a factory that produces products of that general type must do some design work, recalibrate its machinery, build prototypes, test for product safety, possibly order new inventory or expand existing production lines, and market. Depending on the product, these exigencies may give the original producer enough lead time to avoid market failure. Indeed, the pace of technological development in the United States, notwithstanding the absence of either an industrial design law or utility model law, suggests that there is no general market failure with regard to these kinds of products.

Examination of other products also suggests that bearing know-how "on their face" does not fully capture whether they are vulnerable to misappropriation. Semiconductor chips, for example, do not in any meaningful sense bear their know-how "on their face," but a skilled semiconductor manufacturer can discover a new chip's design, apparently much more cheaply than designing a functionally identical chip from scratch, and get a competing chip into production before the first creator has had the opportunity to recoup costs at a fair return on investment. Products based on DNA sequences, at least at present, contain much more know-how in methodologies of production than in sequence alone, and even the sequence information may often not be discernible from the ultimate product produced from knowledge of the sequence. So, there is no correlation between the degree to which products actually do "bear their know-how on their face" and the degree to which they are vulnerable to misappropriation. Therefore, attention should be focused on the basic *2603 question of the existence of "market failure" (vulnerability to misappropriation through cheap and easy methods of copying) and its degree. [FN25]

Nor should we expect to avoid "cycles of over- and underprotection" as new products of technology continue to emerge. [FN26] Different products will vary widely in the degree (if any) to which they are vulnerable to misappropriation. Whatever definition of misappropriation is adopted, courts ultimately will have to give meaning to the concepts and terms employed on a case-by-case basis. Inevitably, generalist judges dealing with new products will over- or undershoot for a time until the legal community has fully worked out the optimal balance of underlying policies. That these cycles occur in the judicial determination of the optimal level of copyright protection for computer programs [FN27] is neither a surprise nor in itself evidence that copyright protection of program code is insufficient protection for software. [FN28]

Rather, it seems more plausible that historical cycles of over- and underprotection in application of the sui generis statutes result more from attempts to cure specific market failure problems in terms of the type of work protected rather than under a general paradigm, such as a generally applicable antimisappropriation statute. Statutes protecting functional industrial design do not work, in general, because they are in fundamental conflict with patent law. [FN29] Specific antimisappropriation statutes like *2604 the Semiconductor Chip Protection Act may be eclipsed by technological developments. The cycles of over- and underprotection, to the extent they have occurred, are the result of focusing on the wrong problem. The key to a general solution, if one exists, is to recognize that market failure is the problem and to devise a statutory scheme directly addressing that problem. Although the two primary papers to this Symposium succeed at the former, they fail at the latter: this is partly because of their reliance on necessarily inadequate and usually misleading catchphrases, but primarily because of their focus on specific objects of protection rather than methods of copying that create market failure.

B. Misappropriation as a Third Paradigm

A third paradigm could have several important advantages that might justify the efforts and risks involved in seeking new legislation. First, it could expressly deal with misappropriation, or market failure, thereby alerting courts to the underlying policies involved from the outset. In addition, concretely defining the basic provisions of the protective scheme would reduce the degree of over- and undershooting as courts struggle to balance the conflicting policies in close cases. Finally, if the scheme is general enough to cover not only those market failures we can now identify but also those that arise with future technological developments, it could reduce political wrangling over the proper degree of intellectual property protection that should be afforded. The question is whether a third paradigm, in the form of a general antimisappropriation statute, is, in fact, feasible.

The feasibility of creating a general antimisappropriation statute depends on whether and how we can answer some basic questions. The most important are how we identify the "market failures" that justify bringing a particular method of copying under the new regulatory scheme, the products to which the regulation relates, and the action we require to decide these issues. Even if a general operational definition of market failure is discovered, however, many details remain to be worked out. We must ask whether these details will be applicable to all products subject to market failure or whether we need a series of sui generis schemes, each tailored to fit the needs of the particular incidents of market failure. These details involve the period and scope of protection, whether independent development infringes, and whether injunctions will be available and under what circumstances. Some of these details, such as independent development, injunctions, and (to some extent) the scope of protection have fairly clear answers when the analysis focuses on methods of copying instead of specific objects of protection. Others, like the term of protection and the operational definition of "market failure" itself, require more analysis, and perhaps more empirical data, than this short Comment on the two primary papers can supply.

1. Objects of Protection Versus Definition of "Market Failure" -- It is clearly tempting to define "objects of protection," such as databases, computer *2605 programs, or program algorithms, as the focus of the new protective paradigm, as the two primary papers attempt to do. In general, this is the approach taken under copyright as well as the sui generis schemes described by Professor Reichman. [FN30] The problems are whether every object placed under the new umbrella truly presents a case of market failure and whether the scheme goes no farther than necessary to rectify the market failure. Including "nonselective databases," for example, runs the risk of covering the factual information they contain (independent of the method by which the information is taken), contrary to our longstanding and well-founded reluctance to protect facts under intellectual property law. [FN31] Including "program algorithms" similarly runs the risk of covering unpatented functionality, independent of the method by which the algorithm is taken from the program, again contrary to our longstanding and well-founded reluctance to protect functionality, as such, outside of patent.

Professor Reichman refers to the objects of his protective scheme as "legal hybrids," but the policy basis for affording them protection is that they are vulnerable to misappropriation -- cheap and easy copying before their original creators have a fair opportunity to recoup investment costs. The term "legal hybrid" is not self-defining, however, and the notion that the objects of protection "bear their know-how on their face" is deficient in bringing out the underlying policy of avoiding market failure. [FN32] As pointed out earlier, some of the most likely current examples of market failure do not involve know-how at all. Rather, they are valuable "sweat of the brow" works like nonselective databases and digitally formatted public domain texts. Such works were, of course, taken out from under the copyright umbrella by the Supreme Court in Feist Publications, Inc. v. Rural Telephone Service Co. [FN33] Others, like semiconductor chips, program design elements, or DNA- sequence information, while functional, do not bear their know-how "on their face" in any meaningful sense, although a skilled competitor may be able to undercut price through relatively cheap and rapid copying. Therefore, any antimisappropriation statute that implements a general third paradigm must focus on the evil to be avoided -- misappropriation, or "market failure" -- rather than on the nature of the works protected.

The general concept of market failure, or vulnerability to misappropriation, seems relatively clear: We must identify those situations in which cheap and easy copying possibilities -- "trivial acquisition of behavioral equivalence" in the words of the Manifesto [FN34] -- threaten to raise a disincentive to the creation of socially desirable works. A major question is whether the definition can be made more precise. I am skeptical that *2606 we will be able to do much better, in view of the historical difficulties with so many works that may well have very little in common except their vulnerability to misappropriation.

One possibility is to couch the definition in general conceptual language and to allow the courts to apply the standard on a case-by-case basis. This would essentially generalize and revivify under a different name the "sweat of the brow" movement that the pre- Feist courts were experimenting with under copyright. Predictability would not be the outstanding feature of such a regime, at least in its early phases, but vague terms are not new in intellectual property statutes, and relatively concrete understanding could well emerge with time. [FN35] In particular, because the statute would make clear that only unfair methods of copying were unlawful, the danger would be reduced that courts would expand the scope of protection to cover otherwise unprotected elements of the works, such as facts.

Another approach would be to state the conceptual standard (market failure) statutorily along with either an exclusive or a nonexclusive set of examples. An exclusive list would not leave room for market failures discovered in the future (absent further statutory amendment), while the nonexclusive list brings back some of the unpredictability inherent in a free-wheeling judicial approach. Possibly an administrative agency could be created that would be empowered to investigate products subject to market failure, hold hearings, and promulgate new additions to the protected class. This carries with it the danger that the agency would become a captive of special interest groups favoring more protection than is socially desirable. In the intellectual property arena, organized interests favoring protection often seem to have more weight in the political process than the diffuse interests of the general public. These organized interests usually have a direct financial stake in legislative or regulatory change, while the countervailing public interest is the amorphous goal of advancing technology and culture through enhancement of the public domain and the incremental improvement of existing works.

Considerable thought, therefore, must be devoted to the problem of relating the concept of misappropriation to the actual objects of protection and the methods of copying to which they are vulnerable. The presumption of a new paradigm should be that, when a product falls outside the protection of the two major paradigms, it falls outside for valid social policy reasons that have been honed and refined over years of statutory tinkering and judicial interpretation. The burden of demonstrating a specific example of market failure, as well as the degree of market failure, should be placed on the forces seeking protection, regardless of whether a court, an administrative agency, or Congress itself is making the decision. *2607 In this connection, I respectfully suggest that the authors of the Manifesto have not demonstrated a sufficient degree of market failure with respect to any noncode aspects of computer software to justify their present inclusion under a general regime of antimisappropriation protection. The kind and degree of empirical economic evidence that should be required to carry the burden of justification for inclusion is another question, however, that requires much more investigation and debate.

2. Term of Protection -- Any concrete scheme of statutory protection for works subject to market failure must come to grips with a definitive term or period of protection. Again, it remains a serious question whether a "one size fits all" approach can be taken for all works vulnerable to misappropriation. A period of protection comparable to that of copyright against misappropriative and verbatim copying of nonselective electronic databases does not seem inappropriate, for example, whereas protection of the functional outputs of a computer program (the user interface) for even five years may jeopardize the benefits society derives from incremental improvements in industrial products. The term of protection must therefore be derived from the market failure analysis that qualifies the work for inclusion within the new scheme. The key question would be how long a particular type of work must be protected to give its creator a fair chance for a return on her investment. Much work remains to be carried out in developing standards for analyzing and resolving this question.

3. Scope of Protection -- The scope of protection for the newly protected works is a problem that must be faced in any statutory scheme. [FN36] The Manifesto would determine the scope of protection by looking generally to the relative size and manner of the appropriation, whether there was an improvement, the degree of similarity, and degree of proximity between the markets. [FN37] All lawyers know how to make litigation arguments in terms of "all the facts and circumstances" based on lists of factors like this. Lawyers also know, however, that these lists of factors are nearly always useless for planning purposes; predictability of results is the inevitable victim of defining liability in such terms. Moreover, these particular factors derive largely from copyright reasoning. They may be appropriate for nonfunctional works like databases that are vulnerable to misappropriation, but not for functional works, where looking at degrees of similarity or proximity of markets essentially ignores the incremental nature of technological advance that has led to the fundamental differences between patent and copyright in the first place.
Part of the answer to the scope-of-protection problem may come from the focus of antimisappropriation protection on the manner of the appropriation as opposed to the substantive content of the taking. It may *2608 be better to abandon the term "scope of protection" altogether, coming as it does with so much historical baggage associated with patent and copyright. Presumably, the market failure analysis that qualifies something for antimisappropriation protection will also identify the methods of taking that need to be regulated. Thus, for any given product, the methods of copying against which it is protected (i.e., the scope of protection) should be reasonably clear. Even so, we are faced with problems similar to those encountered in defining the protected classes: Can a single regime protect all present and future market failures or do we need a series of sui generis regimes?

4. Independent Development -- Neither Professor Reichman nor the authors of the Manifesto are very clear about whether independent development of an otherwise infringing work will result in liability under the new scheme. In principle, however, if infringement depends only on a misappropriative manner of taking rather than on the substance of what is taken, as suggested herein, independent development becomes a clear defense: Anything created independently cannot involve a misappropriative taking. The primary legal problem is allocating the burden of proof with respect to this conceptually easy, but often factually difficult, concept. Again, suggestions on how to effect this allocation in specific cases should be apparent from the economic analysis that qualifies a particular product for antimisappropriation protection against particular methods of copying.
5. Injunctions -- Similarly, neither Professor Reichman nor the authors of the Manifesto are clear about whether injunctions would be permitted under the new regime or whether the rightholders must find their remedy solely in damages (in effect, a compulsory license). Both papers purport to abjure "exclusive rights" systems, which seems to imply that injunctions would not be available, but both also refer to "blocking periods" after a protected product comes on the market. Focusing on the wrongfulness of the manner of taking rather than on the substance of the taking again helps to resolve the problem. There would be no liability at all under the new regime, and therefore no injunctions, for takings that are not misappropriative. Injunctions against any use of misappropriative takings, however, are precisely within the goals of the new regime, at least for the period the competitor would have needed to acquire the information taken by legitimate means.


The two primary papers presented in this Symposium have made an important contribution by identifying "market failure" as central to the ongoing dialogue on how to optimize intellectual property rights in products falling through the crack between patent and copyright. The task now remains to develop the actual legal edifice that will effect this new mode of protection. That process must stay firmly focused on the notion of market failure (misappropriation) -- cheap and easy methods of copying *2609 in relation to the cost of original development -- in defining the products protected, the term of protection, the availability of injunctions, and the activities that constitute infringement of the antimisappropriation right. The key to keeping the analysis properly focused lies in remembering that neither strong similarity nor even complete identity of the final products should, in itself, lead to a finding of infringement of the antimisappropriation right. Rather, infringement determinations must be based on misappropriative methods of copying that, if permitted, would too severely undercut the incentives to create products vulnerable to such methods.

[FNa]. Professor of Law, Arizona State University. B.S.E. 1961, Princeton University; M.S. 1963, Ph.D. 1965, University of Illinois (Urbana); J.D. 1972, University of California (Berkeley).

[FN1]. J.H. Reichman, Legal Hybrids Between the Patents and Copyright Paradigms, 94 Colum. L. Rev. 2432 (1994) [hereinafter Legal Hybrids].

[FN2]. Pamela Samuelson et al., A Manifesto Concerning the Legal Protection of Computer Programs, 94 Colum. L. Rev. 2308 (1994) [hereinafter Manifesto].

[FN3]. See, e.g., Legal Hybrids, supra note 1, at 2442; Manifesto, supra note 2, at 2314 n.14, 2340. The Manifesto also refers to overprotection as a potential market failure. See Manifesto, supra note 2, at 2311 n.5, 2414. However, the emphasis is on underprotection resulting from "trivial acquisition of functional equivalence." See, e.g., Manifesto, supra note 2, at 2340.

[FN4]. For example, products that "bear their know-how on their face," "industrial compilations," "program behavior," and "modified trade secret[s]." See infra notes 25 29 and accompanying text.

[FN5]. See Legal Hybrids, supra note 1, at 2454 2503.

[FN6]. By "nonselective database" I mean essentially any electronic database that fails to qualify for copyright protection due to the absence of creativity in either selection or arrangement. See Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 348 50, 357 59 (1991). The LEXIS and Westlaw databases of judicial opinions are nonselective, for example, because they include all opinions issued in the United States. If that choice were considered creatively expressive, no one else would be permitted to create a similarly complete database of judicial opinions.

[FN7]. There is nothing "hybrid" about a database. It is simply copyright subject matter that must qualify for protection as a compilation. Unless its maker demonstrates creativity in its arrangement or in the selection of its contents, copyright protection cannot be applied. See 17 U.S.C. s 101 (1988) (definition of "compilation"); Feist, 499 U.S. at 348 50, 357 59.

[FN8]. See Legal Hybrids, supra note 1, at 2535 37.

[FN9]. See id. at 2451 52.

[FN10]. See Brief Amicus Curiae of Professor Dennis S. Karjala & Professor Peter S. Menell at 8 n.8, Lotus Dev. Corp. v. Borland Int'l, Inc., No. 93 -2214 (1st Cir. filed Dec. 14, 1993); Dennis S. Karjala, Copyright Protection of Computer Programs, Reverse Engineering, and Professor Miller, 19 U. Dayton L. Rev. 975, 976 - 83 (1994) (symposium issue) [hereinafter Reverse Engineering and Professor Miller].
The definition of "functionality" is central to this claim. Copyright does, of course, protect many works that are "useful" to human beings. Maps enable us to go from one place to another; recipes tell us how to bake cakes; accounting books explain how to implement a particular system of accounting; pictures are useful for decorating homes and offices. All of these works are copyright protected. They are not, however, "useful articles" within the definition in the Copyright Act: "A 'useful article' is an article having an intrinsic utilitarian function that is not merely to portray the appearance of the article or to convey information...." 17 U.S.C. s 101. This definition was adopted in an effort to deal with the problem of industrial design by excluding utilitarian functions other than to inform or to portray an appearance from copyright protection and even excluding expressive aspects of useful articles to the extent they are not separable from the utilitarian functions. See 17 U.S.C. s 101 (Supp. 1994) (definition of "pictorial, graphic, and sculptural works"). However, it represents a statutory description of the kind of "usefulness" or "functionality" that has always been excluded from copyright protection, as opposed to the "usefulness" inherent in maps and recipe books that have long been a part of copyright.

[FN11]. See Dennis S. Karjala, Copyright, Computer Software, and the New Protectionism, 28 Jurimetrics J. 33, 39 (1987) [hereinafter New Protectionism].

[FN12]. See Legal Hybrids, supra note 1, at 2439, 2440 41.

[FN13]. See International News Serv. v. Associated Press, 248 U.S. 215 (1918).

[FN14]. See Fla. Stat. Ann. s 559.94 (West 1988) (repealed 1991); Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989).

[FN15]. "Trivial acquisition" is defined as producing an imitation with "only modest effort, time, and cost by comparison to the resources required to develop the [innovative] product initially." Manifesto, supra note 2, at 2337. Reverse engineering processes are acceptable today because they require more than nontrivial effort, but the automated tools of future technology may make them unacceptable tomorrow. See id. at 2341 42. Unfortunately, the Manifesto itself misses the importance of this insight by later concluding that the value of the know-how itself should be the focus of regulation, rather than particular methods of its acquisition. See id. at 2342.

[FN16]. See Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 352 57 (1991). See generally Dennis S. Karjala, Copyright and Misappropriation, 17 U. Dayton L. Rev. 885, 899, 907 (1992) [hereinafter Copyright and Misappropriation] (pointing out that a broad reading of Feist would leave a wide variety of works vulnerable to misappropriation but that even a narrow reading cannot save nonselective electronic databases to the extent they are regarded as factual compilations).

[FN17]. I have argued that we might stretch copyright a bit to cover this situation by recognizing originality in the method of fixation, an approach that results in very thin copyright protection and is equally applicable to exact art reproductions, photographs of public domain objects, and new typographic arrangements of public domain texts. See Copyright and Misappropriation, supra note 16, at 900 09. The Copyright Act's definition of "compilation" in 17 U.S.C. s 101, however, together with Feist, 399 U.S. at 340, prohibits extension of this theory to nonselective databases.

[FN18]. See New Protectionism, supra note 11; Reverse Engineering and Professor Miller, supra note 10; Dennis S. Karjala, Recent United States and International Developments in Software Protection, 16 Eur. Intell. Prop. Rev. 13, 58 (1994) [hereinafter Recent Developments].

[FN19]. 804 F. Supp. 337 (M.D. Ga. 1992).

[FN20]. See Recent Developments, supra note 18, at 16, 60 61.

[FN21]. The statute defines a computer program as "a set of statements or instructions." 17 U.S.C. s 101 (Supp. 1994). The most natural reading of this language is that it refers to program code, which is the only aspect of computer software that nearly everyone agrees is currently the subject of market failure. See Dennis S. Karjala, Copyright Protection of Computer Software in the United States and Japan, 13 Eur. Intell. Prop. Rev. 195, 197, 231 (1991); Reverse Engineering and Professor Miller, supra note 10, at 988 89. Courts still have not generally recognized the implications of this definition for the scope of protection in a program copyright, that is, that copyright protection in a program should be limited to program code and electronic or mechanical translations of program code. Although Altai cut back on the overly broad scope of protection afforded by earlier courts to computer programs, even that decision continues to recognize the theoretical possibility that protection extends to "non-literal components" of the program. See Computer Assoc. Int'l, Inc. v. Altai, 982 F.2d 693, 701 03 (2nd Cir. 1992). If courts were to begin anew to protect abstract program elements like algorithms or structure, sequence, and organization under the program copyright, there is nothing in either the statutory language or the legislative history that would limit protection to demonstrated cases of market failure, because copyright protection of traditional works of authorship is not based on market failure. The result could easily be another round of judicially created protection of noncode functionality under copyright without any showing of market failure.

[FN22]. See supra note 17 and accompanying text.

[FN23]. The design of a boat hull has utilitarian functions that are inseparable from its aesthetics. See 17 U.S.C. s 101 (Supp. 1994) (definition of "pictorial, graphic, and sculptural" works). It is admittedly unclear how the statute could in any event be rewritten to prohibit only misappropriative methods of copying and not other methods that end up with exactly the same functional design. This lack of clarity is not surprising, given that no one has ever solved this age-old functional design problem (as Professor Reichman has so often demonstrated).
A revision of the patent statute to allow the states more room to maneuver than has been recognized by the Supreme Court, Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989), might also alleviate the problem somewhat. There seems little reason in general, however, to think that the states would do any better at protecting against misappropriation than a federal statute, and reliance on the states could result in a patchwork blanket of protection. Copyright and Misappropriation, supra note 16, at 897 99.

[FN24]. See Manifesto, supra note 2, at 2416 17.

[FN25]. The Manifesto itself provides a good example of the problems that the concept of a work bearing its know-how on its face can bring. Conceding that decompilation of a program is at present a "painstaking and time-consuming process," the Manifesto nonetheless concludes that knowledge gleaned through decompilation lies "near the surface" of the product, even though only future technology will make acquisition of this know-how a trivial matter. See Manifesto, supra note 2, at 2367, 2392. Based on this notion that the know-how lies "near the surface," the Manifesto ultimately concludes that underprotection of "industrial design elements of software" is one of the most pressing problems facing software developers, without ever providing a factual foundation for this assertion. See Manifesto, supra note 2, at 2337. The Manifesto is to be commended for its general emphasis on a market-oriented approach. It should therefore adhere to economic analysis rather than use catchphrases essentially assuming the conclusion that new protection is necessary.

[FN26]. See Manifesto, supra note 2, at 2357; Legal Hybrids, supra note 1, at 2519.

[FN27]. See Manifesto, supra note 2, at 2357 61.

[FN28]. Similarly, the term "industrial compilation" immediately calls to mind copyright notions of "compilation," with protection based on creative selection or arrangement, for what is essentially a functional work. We do not protect automobiles as "industrial compilations" of fenders, radios, mirrors, brakes, and speedometers, no matter how creative or popular a particular model happens to be in this respect, and it is unnecessarily confusing as well as potentially misleading to apply the term to collections of program "behaviors." In fact, talking about program "behavior" is itself an unnecessary verbal substitute for program functionality. Given the difficulty the courts had in ultimately recognizing that program functionality, beyond code, should not be copyright protected, it is dangerous to give them a new term with the same meaning on the basis of which they can again go astray.

[FN29]. Patent law deliberately leaves "obvious" functionality in the public domain. Functional design laws attempt to end-run this result by protecting patent-ineligible designs without the balancing public domain safeguards of patent law.

[FN30]. See Legal Hybrids, supra note 1, at 2454 2503.

[FN31]. See, e.g., Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340 (1991).

[FN32]. See supra note 25 and accompanying text.

[FN33]. 499 U.S. at 353 54.

[FN34]. Manifesto, supra note 2, at 2341.

[FN35]. The "idea/expression" distinction and "fair use" defense of copyright are vague general concepts that, over time, have acquired relatively precise meaning in many concrete applications. See, e.g., supra note 10.

[FN36]. The Manifesto addresses, but does not resolve, the much narrower question of how many algorithms one would have to take from a program to infringe under a system that protected "industrial compilations." See Manifesto, supra note 2, at 2383.

[FN37]. See id. at 2411.