Click for next page ( 82


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 81
5 The Open Agenda The aim of this project was to advance the state of knowledge and the quality of the public policy debate on intellectual property pro- tection issues for software by bringing together an array of interest- ed parties. This area of technology remains in rapid transition, and there are not simple solutions to the complex problems it presents. While a broad analytic structure remains a future goal and a worthy objective of public policy research, the two-day forum whose dis- cussions are summarized here was an important step in aiding the communications among the technical and legal experts who often use different vocabularies and have conflicting problem-solving approaches. Software is evolving from a technology originally conceived as a flexible and inexpensive mechanism for controlling computer hardware to products that embody the functional processes and knowledge base of entire industries and dominate the costs of computer usage. Thus the intrinsic value of software, apart from its form of expression, is hard to quantify, but it is rapidly growing and constitutes the asset end users as well as software vendors seek to protect. A snapshot of the current technology shows evidence of the evolu- tion of the field, with different strands overlapping. That picture reveals that part of the enterprise resembles the highly experimental and entrepreneurial situation of 25 years ago, complete with computer hackers, developers producing "freeware," and highly innovative in- dividuals working in isolation. At the same time, however, it also reflects a major industry dedicated to building reliable, well-main 81

OCR for page 81
82 INTELLECTUAL PROPERTY ISSUES IN SOFTWARE tained software for the efficient execution of well-established appli- cations. Software users cover an equally broad range, from millions of novice users to giant corporations whose software expertise rivals (and sometimes exceeds) that of their software suppliers. With as much diversity and heterogeneity as are found in this sector, it should not be a surprise that discussions involving intellectu- al property protection for software should generate great controversy and complicate agreement on a national policy agenda. The ques- tions involved in the debate what to protect, how to protect it, and for how long-call for value judgments in a large community with few shared values. Because software is so malleable, representing information in many forms (images, sounds, data, and words) and providing the means for creating as well as transmuting and transmitting it, there was little agreement except in general terms on how to describe the values embodied in software. Thus there is less agreement on what aspects of software should, in the public interest, be protected as personal or business property. Complicating the discussion further is the absence of unambiguous technical or legal definitions of some of the key terms of the discussion, for example, a software interface. Not only did the technical and legal experts use the word very differently on occasion, but even the experts also used the term slightly differently, depending on their professional perspective or segment of the industry. Software is of great economic and functional value to society. It is important that software protection not be thought of as an end in itself, but rather as a part of the incentive structure leading to the creation, diffusion, and use of software innovations. Differences of opinion over software protection should not be seen as a battle be- tween opposing economic interests a struggle among vendors and between them and their customers. Rather, the pressing issues revolve around the incentives and disincentives that are provided for cre- ativity, for entrepreneurial risk, for quality services to end users, and for a stable, competitive marketplace. Few would deny that the pace of software innovation and the growth of the industry attest to the strength of the incentives and the adequacy of safeguards to date. The question is whether technological and legal developments in the future will combine to enhance or undermine those incentives and safeguards. At the root of the debate about software protection is not the preservation of the property rights of its creators, but the extent to which protection of those rights will promote innovation without retarding technical progress or induc- ing in the market an instability that might deprive software users of good service as well as new capabilities.

OCR for page 81
THE OPEN AGENDA 83 Just as technological change affects the evolution of legal prin- ciples, so also will legal rulings affect the manner in which technical progress unfolds. For technological reasons, progress in computer hardware has been even more rapid than that in software for many years. Thus, in spite of extraordinary industry growth, software costs and complexity have become the pacing factor in the implementation of many worthwhile applications. At issue now is whether legal developments will widen or narrow the software bottleneck to progress. While recognizing that questions arise at the margins of copyright and patent law, several legal experts advised that the current level of uncertainty should not be interpreted as proof of the inadequacy of either body of law. "One of the factors that is operating here," Goldberg suggested, "is perhaps a philosophic discomfort with things or con- cepts that can't be easily pigeon-holed . . . in the fashion of binary digits-being either-or." He continued, "It is the nature of the legal process that it develops. It cannot be handed down from Capitol Hill with pristine purity and crystalline clarity in its application to all circumstances for all time." Indeed, most should not be and some cannot be. Automated re- verse-engineering and recompilation techniques, for example, may be used to create derivative products that offer the same functional- ity as the original software but are so dissimilar in appearance and structure that even the most astute judges and juries will be hard- pressed to identify illegal copying. The answer to that problem may be technological, such as "fingerprinting" or otherwise identifying the source of derivative works. Or it may be the fact that products created this way may be more costly to maintain and evolve than is original software created from scratch. Yet other important ques- tions fall squarely within the realm of the law, or into the gray area between accepted business practices and law. Different views exist on how best to proceed. DEFINING A CONCEPT OF VALUE Lotus's Frank Ingari struck a responsive chord when he asserted the need to define for software a concept of value that can serve as a lens for evaluating intellectual property issues. Too little discussion, he maintained, is devoted to determining what to protect and why. "We are sort of at the second-order discussion already," he explained. "Is copyright better? What is good for the industry? What is good for the Third World? Everybody is taking positions on what is good for whom, and I don't see much discussion of what we are trying to protect in the first place."

OCR for page 81
84 INTELLECTUAL PROPERTY ISSUES IN SOFTWARE Esther Dyson, publisher of "Release 1.0," concurred. "The real issue here is defining intellectual property," she said. "We can work out what the law does if we can define the stuff that we are trying to protect." Dyson advised, however, that the value of different soft- ware components will change as the technology advances, making value an evolving concept. Initially, she said, "virtually the only thing you wanted to protect in software was the code. Then we got into the 'look and feel issue.' As we move on toward object-oriented code, we are going to have these modules of functionality that are specifically designed to be reused. But the people who design them are still going to want to protect them, charge for them, and so we are going to have a much more complicated problem in the future, . . . when bits of software have to work together." Before deferring intellectual property matters to lawyers, it was suggested, software firms may be better served by first determining what elements of software are most beneficial to users, the ultimate judges of value. "I am appalled by the crudity of the discussion right now in the industry around what is interface and what is functional- ity," Ingari said. "I think we should do a whole lot more worrying about the way these expressions and this creativity break out in pieces and components." Establishing a common understanding of software, as both a func- tional, marketable good and as the product of a complex design and engineering process, serves two necessary purposes. First, it provides perspective on where the value, or intellectual property, lies in a particular piece of software and, therefore, on what elements warrant protection. Second, a broadly accepted notion of software-one that em- braces its numerous manifestations and its complex underpinnings- can guide the application of intellectual property law. LEGISLATIVE "SOLUTIONS"? Few would quibble with the goals for the intellectual property system Robert Spinrad suggests (Box 5.1), but most would argue about measures proposed to achieve them, which range, as Mitchell Kapor pointed out, from "doing nothing" to "doing something radical." Most radical of all perhaps are proposals to create a sui generis system, a body of law specific to software. An argument advanced in support of this notion contends that, at least in the copyright area, attempts to address issues related to the functionality of software are distorting the law. As a result, the argument continues, a sui generis system for software is evolving piecemeal through the case law

OCR for page 81
THE OPEN AGENDA 85 BOX 5.1 SUGGESTED GOALS FOR THE INTELLECTUAL PROPERTY SYSTEM While leaving legal specifics for others to debate, Robert Spinrad of Xerox Corp. offered a set of general goals for the intellectual property system to achieve. Comments by forum participants suggested that Spinrad's desiderata encapsulated the essential requirements of innova- tors and investors. His criteria "The Five Cs" are summarized below. Coverage, or protection, should extend to the "brilliant idea" embod- ied in a software product. ''This is certainly something you want to be - able to protect and own and control the future of," Spinrad said. Protec- tion should also be accorded the programming efforts, the "hard work" that transforms the idea into a marketable product. Continuity, "the ability to build on existing standards and conven- tions at reasonable cost," is necessary to create a foundation upon which the software industry can build. "Access, not appropriation [is keyed Spinrad said, acknowledging that "reasonable cost" is not easily defined. Yet, he added, arbitrary rules or constraints should not "force the programs that work on The user's] behalf to use different interfaces, to meet different standards, . . . to follow different protocols. So, there has to be some way of being able to build one brick on top of another." Consistency in the application and scope of intellectual property protections affords the "predictability, the calculability" that firms re- quire to make the marketing and development decisions that dictate the allocation of financial resources and personnel. Surprises, such as those that might result as belated declarations of property rights, compound the unavoidable uncertainties of the marketplace. Cognizance, "the timely awareness of other intellectual property rights claims," minimizes the danger of being blindsided by competitors. ~1 don't want to be put in a position of developing a product, Spinrad explained, Manly to discover a year after it is on the market-or, even more frightening, just as it is about to hit the market that an essential element of it is something that had been percolating through the patent process and that, because of the confidentiality [of the process!, I didn't know it was cam in"." Convenience, or a straightforward intellectual property system that minimizes the need for litigation, may be the equivalent of "asking for the moon," Spinrad said. Nonetheless, he added, "I would like a mini- mum amount of conflict about which set of rules or . . . statutes cover" which aspects of software. without the benefit of the foresight that would go into a deliberately crafted set of laws. Mention the term sui generis at a gathering devoted to software- related legal issues, however, and strenuous objections are sure to

OCR for page 81
86 INTELLECTUAL PROPERTY ISSUES IN SOFTWARE follow. One frequent criticism holds that abandoning a long-stand- ing body of laws albeit laws that did not anticipate the development of software" for an entirely new system would increase uncertainty, not reduce it. "I think what we might be faced with," said Ron Palenski, counsel for the Association of Data Processing Service Organizations (ADAPSO), "is a conflict between a slower societal process to resolve these issues in a more rapidly moving marketplace and more rapidly moving technology. I would submit, even if you went to a different system, you would still find the same problems." Added attorney Ronald Laurie of Irell & Manella, "I buy copy- right, as opposed to intermediate solutions, because the law in this country, whether we like it or not, evolves interstitially in court. And as someone who spends a lot of time in court, I would feel much more comfortable arguing my case by analogy." Moreover, laws drafted to address today's concerns could be rendered obsolete by future advances in technology, which will engender unanticipated legal issues. That most litigators present at the forum prefer evolutionary de- velopment of case law, within the framework of current copyright and patent statutes, reflects more than confidence that the law is sufficiently elastic to fit evolving circumstances. Their preference may also indicate a lower confidence in the legislative process as the alternative. These litigators share with the scientists an awareness of rapid technological change in the industry but tend to opt for a more surprise-free venue for legal change, while the scientists more willing- ly contemplate the sui generis approach. The chorus of objections that greets proposals for new legal approaches shifts attention to the other pole in the range characterized by Kapor- the "do nothing" option. Howard Figueroa espoused this view, not- ing that many of the issues and concerns now being debated were addressed more than 20 years ago, when firms first contemplated "unbundling" software from their hardware products.2 "The bottom line is that we decided to invest in software develop- ment as a separate business," Figueroa explained, "and we based our decision to put significant resources into that business on the expec- tation that we could protect the expression in our programs from copying. "Copyright protection would apply to that product expres- sion per se. It would require no up-front expenditure to obtain that protection, and it was and is international in scope. We felt that we could build a separate business on this type of protection." Figueroa maintained that the existing intellectual property system works, and he recommended a hands-off approach toward copyright

OCR for page 81
THE OPEN AGENDA 87 law. "I fundamentally believe," he said, "that there is no reason to make any specific changes in relationship to the copyright law as it is currently constituted.... I think that what we should be relying upon is the evolution of the law as it is interpreted in the courts...." While many in the software industry favor staying within the exist- ing legal structure, a substantial number believe that the system should provide better guidance and that copyright and patent laws could be applied more coherently. Thus it is the vast middle ground between the extremes of a do-nothing approach and a sui generis system where most of the discussion and most of the disagreement occur. Kapor was among several technical experts who stressed the need at least to reassess the appropriateness of maintaining the status quo. "[Slaying it is okay to do nothing because things will work out," he said, "strikes me as analogous to the response given by the man who jumped off the top of the Empire State Building. When asked as he passed the 50th floor, 'How's it going?', he replied, 'So far, terrific.' So cautious, thoroughgoing inquiry seems to me to be really justified here." HYBRID SYSTEM FOR HYBRID TECHNOLOGIES? Perhaps neither set of principles is appropriate for software at least not in their current form. This view is espoused by Vanderbilt University law professor Jerome Reichman, who distinguishes be- tween the mature copyright paradigm of artistic property law and a modified copyright approach better suited to what he calls "interme- diate technologies" falling below the patent and copyright paradigms (Reichman, 1989~. His review of international intellectual property laws leads him to conclude that software is the most recent manifesta- tion of "hybrid technologies" that reside in the murky region be- tween patentable inventions and copyrightable creative works. The intellectual property system, he said, carves the universe of created works into art, the province of copyright law, and inventions, the province of patent law. But software, like industrial designs and architectural and engineering drawings, embodies properties of both categories and, therefore, distorts the tenets of patent and copyright law. Historically, nations have differed in their legal treatment of these hybrid technologies, placing them in one category or the other and sometimes oscillating between categories. The results, Reichman said, have never been satisfactory, generating a "cycle of overprotec- tion and underprotection." Problems have been most acute for industrial designs and works of applied art, which are governed concurrently by the copyright and

OCR for page 81
88 INTELLECTUAL PROPERTY ISSUES IN SOFTWARE the industrial property treaties. According to Reichman, both indus- trial art (i.e., designs) and industrial literature (i.e., software) "bear technological and applied scientific know-how on their face." The ingenuity and skilled effort that go into making software and other design-intensive works cannot be hidden, he said. Rather, they are transferred along with the product. Recognizing this vulnerability to easy duplication, many nations have chosen to protect these works through copyright law because, for one reason, it does not discriminate on the basis of merit. According to Reichman, copyright law accepts all comers and it allows the mar- ket to determine value, the very opposite of patent law, which requires the patent examiner to determine merit according to the novelty and nonobviousness standards. From a behavioral standpoint, investors in applied scientific know-how find copyright attractive because of its inherent disposition to supply artificial lead time to all comers without regard to innovative merit (Reichman, 1991~. While the pro- tection afforded by the full copyright paradigm attracts investors in applied scientific know-how, Reichman finds that this paradigm be- comes counterproductive over time because its wide protective net soon frustrates the very incremental innovation that sought copy- right protection in the first instance. Efforts to protect design-intensive works under patent law are un- dermined by that law's standards of nonobviousness and novelty. Although the works are largely utilitarian and sold in markets for non- artistic and nonliterary products (properties that align them with pat- ented inventions), they usually embody incremental improvements on known solutions, Reichman said. Incremental innovation is le- gally obvious by definition and therefore is unprotectable. To grant patents for works that do not attain the high level of originality re- quired for other technologies is to disrupt the patent system, he said. Thus Reichman proposes that a modified copyright approach is more appropriate for applied scientific expertise, including software, and other hybrid technologies than either the mature copyright or patent models that underlie the world's intellectual property system as it stands. He underscores the need to reckon with what he believes is a potentially serious problem, the danger of unreasonably long periods of protection for useful works and the consequent danger of disrupting market competition. Under copyright law, owners of soft- ware innovations, which have an essentially machine-like utility, would hold exclusive rights for up to 100 years. ''No industrial property, no innovation- whether a computer program or a cancer cure or a gene splice should conceivably be protected for 75 to 100 years on the products market," Reichman said.

OCR for page 81
THE OPEN AGENDA 89 Although protection for the better part of a century does seem excessive for any utilitarian product, the discomfort may derive more from appearances than actual harm. Given the rate of progress in computer technology, it is hard to imagine any piece of utilitarian software having market value after 10 or 20 years at most. In any case if only the expression is protected, and not the function, the independent implementation of that function in new software need not be seriously impeded by an overlong period of copyright protec- tion. And all users of the software can experience and learn from the functional utility of a program. An approach to correcting the deficien- cies that Reichman perceives in the legal treatment of software and other design-intensive works is to create a third category of intellec- tual property for hybrid technologies. Echoing Reichman's reasoning, L. Thorne McCarty, a professor of computer science and law at Rutgers University, suggested that soft- ware may represent a new type of intellectual effort. The skilled effort that software development requires, he said, "does not rise to patent-level protection, not on obvious things." Yet, McCarty added, copyright in its current form might not offer sufficient protection against reverse engineering methods that make it increasingly diffi- cult to distinguish between copying and independent innovation. INCREMENTAL IMPROVEMENTS TO PATENT SYSTEM7 Defenders of the current system counter that it is far better to adapt known approaches evolutionarily than to create a new category of protection with all the attendant uncertainties that would ensue. Although there is sharp disagreement over the appropriateness of patent protection for software, even its most ardent advocates find the current patent system to be deficient in some structural and ad- ministrative areas. Donald Chisum of the University of Washington, a strong proponent of patents for software, listed problems in the procurement and enforcement of patents, none of them peculiar to software. While Chisum sees these problems as affecting all tech- nologies, others view the consequences as more severe for software. The need to improve and speed the Patent and Trademark Office's handling of applications is often cited as one such concern. Chisum noted that virtually all other countries publish claims within 18 months of the filing date. In the United States, the patent approval process takes nearly twice as long, greatly increasing the chance that an inno- vator will be blindsided by a competitor whose application was filed earlier. The danger is substantially increased by the prevailing prac- tice of shipping software products in object code. A notification of

OCR for page 81
So INTELLECTUAL PROPERTY ISSUES IN SOFTWARE "patents pending" is of little assistance to a competitor who is unable to reverse compile the product and understand it. Participants also complained of the difficulty of tracking existing software patents, a problem that also increases the risk of unintention- al infringement. The dangers of pursuing an innovation that may already be patented or that may be in the patent-review pipeline are i' of grave concern to developers. According to Dyson, publisher of "Release l.O,J' the first concern could be assuaged by a "meaningful, automatically updating electronic database that contains information on patented technologies." Currently, searches of patent literature are error-prone exercises, ncurring a high risk of overlooking relevant subject matter because of the fragmented, disorganized state of patent information. To cor- rect this shortcoming, Esther Dyson proposed that the Patent and Trademark Office create a database that describes existing software, which would help innovators determine the "prior art" in the field.3 The second issue, that of losing out to a competitor whose applica- tion was submitted earlier, is more problematic because of the short life cycle of software products. Thus the competitor who loses out on a patent has the option of licensing the innovation from the patent holder, if that option exists, or of foregoing the next generation of the product-development cycle. Eventually, warns Brian Kahin, the rapid rate of innovation in the software industry will be slowed to conform with the pace of the patent review and approval process. Finally, several participants advocated establishment of an indus- trial advisory board to help Patent and Trademark Office personnel improve their expertise in the software area. Those who question the appropriateness of patents for software-related inventions include these shortcomings in their appraisals, but only as a starting mint that leads to more fundamental concerns. --I r Even the most ardent advocates of software patents acknowledge that patents issued for broad ideas pose potential problems for the industry. Kapor, who counted himself as neither proponent nor op- ponent, believes this to be an especially serious problem, the impact of which has yet to be felt. A single patent decision that affirms protection for a broad idea, Kapor said, "could change the industry mood from sanguinity to terror." NEXT STEPS The forum discussion reflected a high degree of discomfort by many computer scientists with the intellectual basis for the protection system as it exists. There was considerable uncertainty about how

OCR for page 81
THE OPEN AGENDA 91 well the system will cope with a wide variety of emerging issues. Many features of the present system such as the duration of protec- tion for utilitarian software are recognized as anachronistic. But few would argue that a demand for royalties for use of a 1970-vin- tage word processor would generate any revenue in 2045. Others were critical of the incremental, somewhat stochastic evolution of legal principles based on case law, but no one presented an attractive proposal for near-term legislative action. Nor did anyone advance evidence that the- system the United States has lived with for 30 years has thwarted innovation or failed to produce a business environ- ment that supports rapid growth. Thus considerable common ground united the participants. Nevertheless, there is much work to be done by technically quali- fied experts who understand the underlying legal principles and pol- icy issues. Legal scholars will have to continue to pursue the questions that fall within their domain. But the Computer Science and Telecom- munications Board, although not competent in the law, has access to a broad range of technical experts, many of whom have been deeply engaged in the legal and political dimensions of their trade. These suggestions for further exploration are addressed to the board. When the steering committee began preparing for the forum project early in 1989, it was struck by the paucity of scholarly literature on the nature of software, the values it may embody, and the balance of creativity, discipline, structure, and knowledge of applications that underpin those values and lead to excellent products. Copyright protection has operated in the software market during many years when most system software was unique to the hardware it ran on and when there were dozens of word processors, accounting programs, spreadsheets, and device controllers of more or less equivalent merit. Copyright is hospitable to genius and mediocrity alike. End users, not patent examiners, judge the social merit of copyrighted software. Bad programs enjoy the same protection as excellent ones, but so long as it is only their expression that is protected, they are discarded without harm to the industry. The market, not the privilege of limited monopoly, has driven the industry's growth. But the magnitude of the assets required to launch a successful software project is growing very rapidly, and the constraints on substi- tution of independent implementations of needed function are rising, too. Barriers to product substitution are rising as industry standards aimed at increased interoperability gain acceptance and end users insist on familiar interfaces. The qualities that are associated with successful products will depend increasingly on collections of talent of a quite unique kind. The incentive to imitate, if not copy, the

OCR for page 81
92 INTELLECTUAL PROPERTY ISSUES IN SOFTWARE work of successful software providers will grow. Whether the at- tributes of patentable invention have their equivalents in software is a matter for debate. But the clamor for incentives to nurture that talent is sure to rise. How then is the rationale for awards of limited monopoly aimed at "advancing the useful arts and sciences" of soft- ware to be developed in the absence of a deeper understanding of those unique talents and qualities? A second, more practical task for computer scientists and lawyers in collaboration is to develop operationally useful characterizations of software attributes that require legal interpretation. Examples might include system-level and user interfaces; languages; the con- cepts of compilation, Recompilation, and restructuring; the distinction between architecture, design, and implementation; and the distinc- tions between mathematical principles, algorithms, and procedures. These characterizations need to be designed to be adaptable, if not invariant, in the presence of rapid technological change. The value of this effort lies in increased clarity of communication and debate, in court and beyond. A third area needing study is the implications of a rapid increase in the rate of issuance of patents covering functions embodied in software. Because the increase in rate of filings and the courts' seemingly growing receptivity to the legitimacy of software patents are rela- tively recent events, it is not possible to point to past growth of the software industry as proof that this growth will not be inhibited or the market disrupted in the future. There are both abstract and practical problems to be addressed. When are clever mathematical procedures intrinsic properties of nature, like Newton's laws of physics, and thus unpatentable? On the practical side, how is either the patent examiner or the software entrepreneur to determine whether a poten- tially patentable software idea has already been used in available computer code, and is thus part of the prior art? How is the developer of a commercial software offering to be able to ensure that the company's programmers are not reinventing ideas that have been submitted for patents not yet issued? What, indeed, are to be the criteria for "nonobviousness" in software? What is at stake? The future success of American innovators in an industry with clear global leadership will hinge in large measure on developments in software design houses and the marketplace. The courts and the Congress, however, will delimit the playing field on which software developers compete. They will do this through the rules that they do or do not make or alter. Given the American propensity to litigiousness, and the drag on productivity that many

OCR for page 81
THE OPEN AGENDA 93 industries have experienced as a result, the software industry is vul- nerable. The delicacy of the current balance of protection arrange- ments for computer software requires respect, even as the issues are vigorously examined. The open and constructive discussion at the forum and the preparatory workshop have made a useful contribu- tion toward a broader understanding of that balance and its implica- tions. NOTES 1. The forum started a dialogue that should be continued. The Computer Science and Telecommunications Board is exploring options to further address the issue. 2. The "unbundling" of software and service from hardware sales by IBM was undoubtedly a wise business decision that depended for its success on IBM's copy- right in the system software. As described by Howard Figueroa, Thomas Watson, Jr., ascribes the decision in June 1969 to general counsel Burke Marshall's warnings about antitrust exposure. See Watson and Petre (1990). 3. The LEXPAT database can provide easy and rapid access to copies of software patents, if the searcher wishes to examine a particular patent and knows its patent number. Dyson's suggestion differs in that it is proposing a content/area search capability that does not currently exist. Developing the proposed capability would be a major undertaking.

OCR for page 81