Keeping the U.S. Computer Industry Competitive: Systems Integration (1992)

By the end of the 1990s, we will have had five decades of experience with computers. But as with several other major commercial innovations (e.g., automobiles, aviation, telephony, and pharmaceuticals), nearly 50 years of experience do not establish computers as a mature technology; nor do they qualify even the most practiced users as consummate authorities. Many observers believe that the true potential of digital technology in all its embodiments will not be realized until well into the next century. This view was expressed by Alan Perlis, the late Yale University computer scientist, in his remarks at the CSTB's inaugural colloquium on the competitiveness of the U.S. computer industry:

Some might substitute information age for Perlis's computer age, but his themes of exploration and continuing mastery of digital technology are no less relevant. In many ways, systems integration marks a new phase of exploration and, as Perlis put it, domestication of the technology. The hybridization of computers and communication technology will yield a new progeny of applications, perhaps revolutionary capabilities that are not discernible from today's limited vantage point.

But these expectations are not shared by all. Some observers extrapolate from today's experiences with computers and related technology, and they are more circumspect in their outlook on the future. The current generation of technology, they say, has yet to be domesticated, and, for now, the need for new types of applications and information services may be greatly overestimated.

Colloquium participants offered their own assessments of the nation's information future, with many generally endorsing measures to build a nationwide information infrastructure. Some described an advanced infrastructure as essential to the nation's economic development. James Fischer of Andersen Consulting suggested, for example, that as the web of computing and communication networks grows, the number and magnitude of the potential benefits also grow. As an enabler of innovation, integration can confer advantages to a single company, he said, "but it also is the basis to give a group of companies some advantage, and then industries, and ultimately, I would assert, the national economy." Other speakers predicted a similar cascading of benefits in education, health care, government operations, entertainment, and numerous other areas of social, organizational, and personal activity.

In Japan and Europe, government and industry appear to find the prospective payoffs more compelling than do their U.S. counterparts, according to Robert Martin of Bellcore and several other speakers. Indeed, Japan's Ministry of International Trade and Industry has estimated that, by the year 2020, the country's planned information network will undergird a full third of its economy.² Current activities there and in Europe are creating a political and economic environment that "encourages cross-industry cooperation" on planning and building an infrastructure for the information age, Martin said. Japan and the European Economic Community, he added, are "spending more than twice as much per capita as the United States to upgrade their infrastructures. Each has sponsored cooperative, precompetitive research to define the infrastructure."³

While overseas activities may eventually yield returns that could help foreign competitors eclipse the U.S. leadership in systems integration, colloquium participants were most troubled by the nation's complacency about the role of information technology in its future. With the exception of the HPCC program, neither U.S. government nor U.S. industry has contemplated an initiative to harness the evolving technology to bolster the nation's economy, improve the performance of its institutions, and better the lives of its citizens. A study by the National Telecommunications and Information Administration released in October 1991 offers some encouragement, because it both articulates the importance of information infrastructure and considers a range of relevant issues from a public policy perspective.⁴ The issues involved are numerous and complex, and many, such as concerns

about personal privacy, cannot be resolved solely in the realms of technology and business.

Because it was intended as a forum for discussion of varying perspectives, the colloquium did not strive for consensus or sharply focused recommendations. Nevertheless, it is possible to distill from participants' remarks several major issues and concerns requiring the thoughtful attention of business, universities, and government. Those include the need for making strategic investments, including demonstration projects by the federal government; attending to technical standards; enhancing the role of universities; pursuing enabling technologies; attending to human elements; providing for system security and privacy; and developing a shared vision of the information age. These themes are recapitulated below.

In 1934, the federal government made universal telephone access a national goal and then built the regulatory framework and provided the incentives that led to the creation of the world's best telecommunications system. Colloquium participants suggested that it is now time to augment that goal in light of technological changes profoundly altering the nature of communication and information. Many also suggested that the federal government can play an instrumental role in leading the United States into the information age.

"What we all want to see the government do," said Mischa Schwartz of Columbia University, "is provide a seed or stimulus to progress in some way. That is what is happening in Japan and in the European countries."

According to Schwartz and others, the short-term, market-driven perspectives of most U.S. businesses will likely translate into a continuation of ad hoc networking initiatives in the private sector, but the sum of these fragmented efforts will not be a cohesive nationwide information infrastructure. Moreover, colloquium participants questioned whether, on their own, U.S. companies will make the long-term commitment necessary to resolve the many technical issues that a nationwide information network poses. An infrastructure will not develop spontaneously, they said, and the decision to build one should be made prospectively, rather than in hindsight.

"Today, . . . I never hear anybody argue about what is best for the United States," said Robert Lucky of AT&T. "It does not enter any equation whatsoever." The nation's telecommunications system, he claimed, is evolving haphazardly, as all activities are dictated by the need to make money. Profit potential has become the "sole source of leadership," he said.

Money, however, is not an incidental item in building a national infor-

mation network, one that links every home and business. Estimates of the cost of installing optical fiber and associated equipment throughout the United States range from $200 billion to about $1 trillion.⁵ Many believe the investment will pay for itself, yielding returns in a variety of ways. For example, one recent study estimated that the ability to communicate, shop, and conduct business electronically over a nationwide fiber-optic network would permit 6 million people to work at home, replace 13 million business trips with teleconferences, eliminate 3 billion shopping trips, and reduce truck and delivery miles by 600 million.⁶ The resultant annual savings, realized in the form of reduced fuel consumption, pollution, and commuting time, were estimated to total about $23 billion. While the transportation-communication trade-off has been a subject of discussion, speculation, and disappointment for decades, we appear to be at the threshold of some real shifts in behavior as a result of changes in technology, the economy, and shifting views on important topics like the environment, education, and family.

Because the cost of building an information infrastructure is large and because many of the anticipated benefits might be considered public goods, there have been calls for cost-sharing and risk-sharing between business and government. In addition, the perceived costs and benefits and the magnitude and complexity of the task of building an infrastructure were seen by colloquium participants as impetus for measures to foster cooperation and coordination.

"This is a very long term issue that requires an unusual amount of cooperation across academia, industry, and government," said Martin of Bellcore. In the private sector, he noted, the broadcast, computer, and telecommunications industries "must learn to cooperate if we wish to have an information network." Thus an important role for government is to serve as an intermediary and to create a forum for cross-industry participation.

Martin and other speakers questioned, however, whether a truly coordinated and cooperative effort is possible within today's fragmented regulatory and policy framework. Federal and state policymakers and regulators, he said, "are doing superb jobs within their own domains but, I fear, are looking backwards towards the old edict of universal telephone service . . . . [W]e need to carefully look at how we are stimulating investment in this infrastructure in the United States."⁷

The fact that U.S. companies spend several billion dollars on private networks indicates a need for a communications infrastructure. But does this need extend beyond business, especially large firms and organizations such as federal agencies? Still a matter for debate is the extent to which U.S. consumers may wish to partake of the smorgasbord of information-age

services that becomes possible when households are linked to a nationwide network. Commercial ventures in videotext and other information services (e.g., CompuServe, Prodigy, GEnie) have served relatively small numbers of customers to date. The development of narrowband ISDN and its anticipated growth into broadband ISDN are largely the product of telecommunications carriers' intentions, rather than burgeoning consumer demand and brimming societal expectations. On the other hand, these services do not go as far as those envisioned for the National Research and Education Network (NREN), which is driven by advanced scientific and technological demands plus concern for the needs of education.

Commitment to and momentum toward an advanced broadband network must be generated by new applications that appeal to a broad number of people. Uncertainty about future markets for information services and applications can impede steps that would further the development of a national network. According to several colloquium participants, national and regional demonstration projects are required to evaluate prospective consumer-oriented uses of an information network. Today, several local experiments—at least 50, based on a 1990 estimate⁸—are evaluating household use of fiber optically delivered information and entertainment services. But whether the results of these trials will provide definitive evidence of the appeal and utility of such services on a national scale is doubtful.

Lucky of AT&T noted that with AT&T's divestiture of the Bell Operating Companies, the United States lost its only mechanism for conducting ''pioneering experiments'' to try out new services. "We would give communities certain advance services and see how they played out," he explained. "We cannot do this any more. They [trials] do not make money. . . . In the days before divestiture, we had a great mechanism for doing all this. We had this rate-base business where we effectively taxed the telephone-paying public to pay for what we thought would be good for them. We need a mechanism that [permits] . . . the nation to devote some resources to doing pioneering experiments."

The authority and means to implement such a mechanism would seem to rest with federal and state regulators, several speakers suggested. They contended, however, that federal and state regulators are principally interested in controlling the cost of telecommunication services to consumers. Longer-term infrastructural issues and their implications for the general public tend to rank low on the regulatory agenda. Without consumer-oriented trials of advanced information services, said Martin of Bellcore, conjecture will continue to dominate the debate over a nationwide information infrastructure. Whether hooking up households and offices on a network will yield such benefits as individualized education, better health care as a result of providing access to remotely located medical expertise, and new economic opportunities for declining rural communities can only

be determined from systematic testing, he said. Similarly, results of service trials could either strengthen or weaken arguments that large businesses would be the primary beneficiaries of an information infrastructure, he added. Evidence, either way, would help guide regulators' and policymakers' decisions on financing infrastructural development.

Existing federal government programs hold varying degrees of value as demonstration vehicles. Particularly promising is the NREN. Beginning efforts to build a multigigabit network, David Farber of the University of Pennsylvania suggested, are in themselves an "enabling technology." Added Samuel Fuller of the Digital Equipment Corp., "Here is an ideal project that can only be done in collaboration with industry, and, I believe, the universities can play an important part. It seems to me [NREN] provides a wonderful, concrete example that will stress a number of key areas that will have to be improved if we are really going to be effective in systems integration. . . . It provides a rich area for us to develop methods, do research, and make progress."

As an additional spur to progress, several colloquium participants suggested, the federal government could sponsor development and demonstration projects that would pioneer applications of advanced information technology to the missions of information-intensive agencies, such as the Social Security Administration (SSA), the Internal Revenue Service (IRS), and the Environmental Protection Agency (EPA). Expanding the demonstration potential of federal systems would build on a large base: the federal government has been the largest consumer of information technology in the world, federal needs for systems have long nourished the systems integration business, and today many agencies have significant needs to modernize their processes and supporting systems. Historically, information technology projects for government agencies such as the National Aeronautics and Space Administration and Department of Defense have broken new applications ground.

Beyond focusing research and development on civilian, or commercial, applications of information technology, federal systems demonstration projects are likely to yield tangible dividends in the form of increased effectiveness and efficiency in government operations, as well as improved decision making. Achieving this promise, however, would be challenging, as such major information-intensive agencies as the SSA and the IRS process information on a scale far exceeding that of single companies undertaking systems integration projects (Tax Systems Modernization at the IRS, for example, is expected to involve technology to handle 100 million individual tax returns plus organizational tax filings, at an investment of $8 billion).⁹ For a variety of reasons, information systems at most civilian agencies have not tended to pioneer innovations; they have in recent years lagged the best commercial practice. Nevertheless, initiatives such as GOSIP, for imple-

menting OSI networking standards; FTS-2000, for upgrading voice and data communications (including ISDN implementation); and other federal procurement activities point to the possibilities for a wide range of federal systems to implement and demonstrate aspects of systems integration.

Another option for stimulating investment, suggested by Mark Teflian of Covia, is to provide tax incentives to private organizations that invest in advanced technology needed to modernize the nation's communications infrastructure. To strengthen the "competitiveness of U.S. business here and abroad," Teflian said, the United States must be able "to get past the point of having to deal with the integration problems at the lower level—the physical transport levels of communications networks." Tax incentives might encourage firms to take the next step and move on to higher levels of integration and internetworking, which, he said, would improve business performance and ultimately benefit the economy.

Efforts to develop communication and information technology standards, one commentator has recently written, "are almost always slow, laborious, political, petty, boring, ponderous, thankless, and of the utmost criticality."¹⁰ Such is the nature of standardization, a process that U.S. companies "tend to flay," said Schwartz of Columbia University. "Sometimes we tend to go with it; sometimes we go against it.''

In recent years, however, more and more U.S. companies have come to appreciate the strategic importance of standards. Manufacturers and service providers recognize the market-expanding effects of standards, and users of information technology are wary of proprietary approaches that limit their future options for connectivity and interoperability. Unfortunately, this appreciation is compartmentalized according to industrial domains. Communications carriers attend to their own sphere, manufacturers of computer hardware and software attend to theirs, and user organizations that get involved in the process focus on their specific needs. Even within an industry, interests may be specialized and separate. For example, providers of wireless communication services and firms that operate public switched networks have done little to bridge their two operating domains and create a larger hybridized communications network.¹¹ Thus standard-setting efforts do not fully reflect the convergence and interdependence of technologies and industries wrought by the digitalization of information.

In Europe and Japan, several colloquium participants noted, standards have been made a top priority by both government and industry. The European Economic Community, for example, has created a standards institute to serve as a focal point of standards-making activity. Collaborators there are placing greater emphasis on developing "prenormative" standards that

anticipate technological advances and future networking needs, but yet accommodate refinement and adjustment when technology and services are ready for commercial introduction. In so doing, European companies are strengthening their position in the international standards-making arena, participants suggested.

Participants made several recommendations regarding the development of standards that foster the growth of a national communications and information technology infrastructure and that can compete for international acceptance. First, they said, U.S. government and industry must adopt an integrated, strategic approach to standards making. An important ingredient of such an approach is cooperative precompetitive research, perhaps jointly funded by the federal government and industry. Although speakers did not propose a specific mechanism for coordinating standards making activities, most called on the federal government to play a formative role.

"The federal government," maintained Albert Crawford of American Express, "should be far more proactive in influencing international information standards. Government representatives should solicit input not only from equipment vendors but also from systems integration firms and leading-edge users."

"Standards activity clearly needs to be expedited," Martin added, "but finding the right model that assures ongoing open participation by industry and yet accelerates the process is a very complex matter that requires careful study."

Many speakers expressed concern over the complexity of current and emerging standards, which, in the long run, could undermine internetworking efforts and perhaps slow the adoption of new technology. Citing one example, Alfred Aho of Bellcore said the communications protocols designed to support the Open System Interconnection reference model "are getting bigger, fatter, more complex, and more numerous. Eventually this whole system is going to collapse under its own weight." Aho described standards as "essential" to systems integration and networking on national and global scales, but he stressed the need to simplify standards and to increase the speed and efficiency of the standards-setting process.

An important challenge for researchers is the development of so-called lightweight communications protocols, those that entail less processing and, therefore, take less time to carry out their functions than do today's "heavyweight" versions. However, current analytical models and other methodological tools now used to design communications protocols, explained Hisashi Kobayashi of Princeton University, may not be applicable to efforts addressing this challenge. "There are a whole range of protocol issues that cannot be addressed by the existing methodology," he said.

Moreover, internetworking efforts would benefit significantly from wider adoption of programs for certifying the conformance of products with

industry-accepted standards and from the development of better mathematical tools to improve the reliability of product testing. However, the former, system certification, is the subject of some controversy within industry centered on how best to demonstrate conformance, while the latter, including verification and validation of software, is the subject of vigorous activity and debate in the technical community.

Given the technical challenges that advanced information networking poses and the systems integration industry's need for a highly skilled interdisciplinary work force, one might expect the nation's universities to be playing key roles as performers of research and as trainers of science and engineering personnel. Unfortunately, this expectation does not mesh with reality. Excluding the involvement of some in the young NREN project and a few other isolated examples, universities have not embraced systems integration as an area of formal inquiry.

Because the field encompasses many disciplines, Kobayashi explained, it "does not really fit in the traditional academic disciplines like computer science or electrical engineering. It is not so easy for faculty members and graduate students to focus on systems integration in the broader sense of the term." That an interdisciplinary approach is required, however, was clear from the remarks of industry representatives such as Jeffrey Heller of Electronic Data Systems, Charles Feld of Frito-Lay, and Fischer.

"Much like the difficulties in making software engineering a legitimate academic topic," added Larry E. Druffel of Carnegie Mellon University, "systems integration faces the issues of definition, codification, and theory." Thus despite the perceived need on behalf of industry, few universities have thus far set up graduate-level programs to address systems-integration issues. Irving Wladawsky-Berger of IBM challenged university researchers and their institutions to begin addressing the field's complex problems or "risk falling far behind the state of the art."

While acknowledging that systems integration is struggling for recognition as a legitimate academic topic, university representatives pointed to some practical impediments.¹² One is the lack of private and public funding for the interdisciplinary research programs that are required to address many of the issues in the field. A related obstacle is the high cost of advanced technology. In the communications area, for example, most universities do not have the resources to make the transition from the old model of point-to-point telecommunications to the rapidly evolving model of networking, said Schwartz. The modernization, he said, "takes a lot of know-how, and it takes a lot of equipment. Nobody is providing that know-

how, and nobody is providing that equipment, and so [universities] are falling further and further behind."

Druffel concurred: "Universities seldom have the funds to support complete integration, and most do not have the engineering expertise to build and maintain an operational system."

Heller suggested several prescriptive actions to foster greater university involvement. Consortia with members from government, industry, and academia, he said, offer one means for assembling the "critical mass of resources and talent" needed to address complex systems integration problems. Heller also advised universities to act on their own—"to concentrate on understanding systems integration, providing input to industry, and teaching" the interdisciplinary and technical skills that the field requires. For its part, he added, "business should assist universities in upgrading equipment, share its expertise, and provide input into curriculum development."

Technological innovation yields the raw materials of the information age, while systems integration offers perhaps the best means for making the most productive and widest use of those materials. The competitive advantages reaped by firms such as Frito-Lay and others that have effectively linked information technology, people, business operations, and long-term strategic planning provide compelling evidence for this assertion.

Occasional examples of delays, cost-overruns, and failed integration projects notwithstanding, U.S. firms have pioneered and demonstrated the substantial benefits that can be gained by networking today's computing and communications technology, and they have given a hint of what the future may bring. In growing numbers, foreign firms and governments will seek to capture the competitive economic advantages that systems integration affords. This will mean an expansion of global markets for systems integration services and the underlying base of hardware and software on which information systems are built. In turn, a rapidly growing market will also attract new competitors to the systems integration business.

These developments, already unfolding, could eventually lead to the repetition of a lesson that is all too familiar in the nation's computer and electronics industries. Being first in the development and application of new technology does not guarantee future commercial success. Yet having given birth to information networking, U.S. firms are at least in a position to capitalize on the second generation of networking, during which entire industries and entire economies and societies will go "on line." That these firms will be able to do so, however, is not inevitable. Almost certainly,

foreign firms will attempt to leapfrog their U.S. competitors and cash in on growing opportunities to sell devices, software, and services. The governments of Japan and European nations are collaborating with industry on a wide range of information technology research and development projects, ranging from the manufacture of advanced semiconductor memory chips to demonstrations of new information services to devising security counter-measures and privacy safeguards.

Colloquium participants identified several key areas of technology and research where continuing progress is essential if the U.S. systems integration industry is to maintain its world-leading position.

"We [must] do for systems integration what computer science and computer engineering have done for computing," asserted Aho. "If we do not do this, then I think the kinds of advantages that we currently enjoy will be eroded, and eroded rapidly."

Behind Aho's assertion lies the widely held perception that systems integration is largely an art or advanced craft, rather than a full-fledged discipline built on fundamental scientific, engineering, and design principles. However, the interdisciplinary nature of the field makes meeting this challenge all the more important and all the more difficult. The transformation from art to science requires contributions from a diverse collection of fields, such as mathematics, ergonomics, and business and management science, as well as software engineering and computer science and engineering. Although many problems can be addressed in distinct research domains, the results of investigations will generally be partial solutions to larger and more encompassing issues. As the field's name makes implicit, integration and coordination of research contributions will be critical to building the scientific underpinnings of systems integration.

What is more, there are intrinsic problems in working on large systems. First among these is the inherent difficulty in designing large systems that will incorporate both automated and non-automated (human, business, or physical) elements. No less a challenge is the innate propensity for large systems to change or evolve over time. For example, 40 to 60 percent of the effort in the development of complex software systems typically goes into maintaining—i.e., changing—such systems. ¹³

As suggested in the earlier discussions of the NREN project, the expected emergence of communications and computer networks that permit

users to exchange data at a rate of several billions of bits per second (or greater) spawns an array of technical issues. There is need for both incremental improvements in existing technology and "clean sheet" approaches to networked computing and communications. In hardware, critically important areas include advanced integrated circuits, computer architectures, photonics and photoelectronics, flat panel displays, switching technology, digital signal compression, and many others. Colloquium participants did not dwell on hardware-related issues, but many acknowledged that the future of the U.S. systems integration industry is closely linked to the fortunes of domestic hardware manufacturers. They underscored concerns about the competitiveness of certain segments of the U.S. semiconductor industry and about the so-called commoditization of computers, a trend that favors Japanese firms and other foreign competitors noted for their manufacturing prowess.

Participants devoted more attention to software, which, as Fuller of the Digital Equipment Corp. (DEC) explained, is the primary source of "value-added" in systems integration. In contrast to the uneven fortunes of U.S. hardware manufacturers, the nation's software makers account for over half of world sales of computer programs and related services, having generated revenues of about $63 billion in 1990. ¹⁴ This world-leading position appears to bode well for U.S. systems integrators as they venture into foreign markets. But again, Japan and Europe are making determined efforts to close the software gap, focusing on design and programming methods intended to make software development more efficient and less prone to quality defects than are the processes used by U.S. producers. Japanese designers and programmers have been highly successful in creating reliable, high-quality software incorporated into consumer electronics products, and they, as well as European software companies, have proven their ability to write customized programs for business applications.

Fuller and other colloquium participants suggested that improved methods for large software development projects would help both U.S. software firms and systems integrators distance themselves from the competition. Especially needed, according to Fuller, are better computer-aided software engineering (CASE) tools for building "programming environments for system control, revision control, and the integration testing and verification of large systems." By making these methods a research priority, he added, "I think we can continue to differentiate ourselves" in the world market for systems integration.

Two other critical areas, Fuller said, are the enhancement of multimedia human interfaces and development of standardized techniques for managing database systems. Both are essential elements of efforts to make computer-aided collaboration and information networking easier and more productive.

Collectively, the nation's businesses, government agencies, and other organizations have invested hundreds of billions of dollars in information technology since the 1960s. This large installed base of devices and, in particular, software and databases is alternatively described as boon and bane. With each wave of advance in commercial technology, organizations are finding it increasingly difficult to combine the old with the new, and yet the information and applications embedded in yesterday's investments continue to be strategically important assets. In fact, the value of these assets may actually increase with the new functions and capabilities made possible by state-of-the-art technology, if the two can be made interoperable.

In the same vein, many organizations have a long list of new software applications awaiting development. Progress is slow because each new project generally begins from scratch, even though routines and subroutines that will eventually be incorporated into the final product already reside in existing applications.

Both problems—the one of re-engineering the installed base and the other of building on past software design and programming efforts to create entirely new applications—are widely recognized. Solutions would be especially beneficial to private and public organizations that have invested heavily in information technology, most of which are located in the United States. "By one mechanism or another," Fuller said, "we have to stand on the shoulders of others rather than re-implement from scratch."

While computers are typically viewed as implements of automation, the technology is most effective not as a substitute for workers, but as a complement to the capabilities of humans. However, the computer's potential as a thinking aid, as a tool that fosters creativity and problem solving, and as a means of collaboration is not easily achieved.

Systems integration, said Laszlo Belady, then of the Microelectronics and Computer Technology Corp., "is not purely a technology issue. It is, to a large degree, a human resource issue." But people may be the most unpredictable and, therefore, the most neglected variable in information networking and systems design and implementation.

"I think we are just starting to understand the fact that, really, the computing system has to be driven by the human system," said Michael Taylor of DEC. "So we are starting to think in terms of a new paradigm that says you start with the people, the way those people need to do work."

Several colloquium participants described systems integration as an agent of change. Yet change can be disruptive. The transition to, for example, a leaner organizational structure and a more collaborative work environment may never be realized, as may also be the case for many of the benefits the organization envisioned when it chose to invest in information technology.

Today, however, there are few tools to help ensure that a distributed computing system will enable increased organizational and personal productivity and creativity. Experience provides ample evidence of the importance of training, ease of use, and well-designed user interfaces. But fundamental knowledge of how people think, work, collaborate, and use information technology could lead to significant improvements in the utility and value of the technology.

The global spread of computing and communications networks heightens concerns about security and privacy. Once organizations begin to link up with others outside their own private network, as many are now doing, said Fischer of Andersen Consulting, "the possibilities for security violations go up exponentially." If not addressed with effective countermeasures, colloquium participants noted, increasing vulnerability to security breaches could constrain applications of information systems. Yet, they said, the escalating risk has yet to receive concentrated attention from industry and government.

The implications of this situation were addressed in a recent study by a committee of the CSTB:

Moreover, the increasing ease with which information on individuals can be assembled, correlated, analyzed, and distributed may undermine the right to privacy, explained Mary Shaw, professor of computer science and software engineering at Carnegie Mellon University. She pointed to two areas of concern. First, Shaw said, information that was once difficult to gather is now easily accessed with computers.

"The second factor that is operating," she added, "is that information that has long been public and probably legitimately so, can now be synthesized," resulting in a composite body of information that is qualitatively

different from the raw data from which it was assembled. ''This qualitative effect is sort of denied by the people who claim they are only bringing together information that is already public," she continued. But it creates "tension between business interests that correlate this information . . . and individuals' personal right not to have it correlated, publicized, and distributed."

Although in itself an important economic activity and a potentially large source of export earnings, systems integration may derive most of its importance from its facilitating role, that of enabling organizations and individuals to use information technology most productively in applications ranging from manufacturing to government services to education to entertainment. Recognition of this value is implicit in the coordinated actions by European countries, Japan, and other nations to build regional and national communications and computing networks.

While not intended to garner unanimity of opinion, the colloquium did generate a consensus view on the critical need for this nation to plan for its future in the information age. Throughout the discussion, there were calls for vision, leadership, and even imagination on the part of the federal government.

"Leadership is not a bad thing to expect from your government," said Wladawsky-Berger. "That is very different from industrial policy or control or regulation."

If government can help U.S. industry stake a claim at the frontier of technology and demonstrate the potential returns from exploring and developing that claim, this line of reasoning goes, then the pioneering efforts will attract many firms to settle and compete for a livelihood at the edge of the frontier. A rich cycle of commercial innovation can be expected to follow, yielding benefits for the rest of society.

But leadership without a committed following, one that is willing to shoulder part of the risk of exploration, is a hollow exercise. Firms now calling on the U.S. government to lead the nation into the information age must also share that vision and commit to making it a reality. A recurring question at the colloquium was whether many U.S. firms are willing to take a long-term view and to invest in the future, perhaps at the expense of short-term profits.

A firm can be "market driven with the best of them," Wladawsky-Berger explained, "but if you do not pay a lot of attention to innovation and technology as one of the major forces that will drive you to the future, then you really do not have a future. I do not know how you can talk about innova-

tion and technology, where the payoffs may be more than five years away, if you are not convinced that that is very important—that you either push at the leading edge of technology or you go out of business."

1. Computer Science and Technology Board, National Research Council. 1990. Keeping the U.S. Computer Industry Competitive: Defining the Agenda, National Academy Press, Washington, D.C., p. 66.

2. Gilder, George. 1991. "Into the Telecosm," Harvard Business Review , March–April, p. 158.

3. For example, within the European Community the Research and Development in Advanced Communications technologies for Europe (RACE) program represents a collaborative effort to develop the telecommunications infrastructure required to support future computer networks. The program was initiated in 1988 as a five-year effort with eight nations contributing over half a billion dollars to develop an integrated broadband communications system for high-speed operation. (Blackburn, J. F. 1989. The RACE Program in 1988, Office of Naval Research Europe Report (ONREUR), Washington, D.C., March; or RACE DG XIII-F (Directorate General XIII). 1990. Research and Development in Advanced Communications Technologies in Europe. RACE '90, European Commission, Brussels, March.)

Japan, for its part, has committed more than $130 billion through the MITI New Media Community and the NTT Information Network System project to building a national ISDN network and achieving full digitization of Japanese telephone service before the turn of the century (Bellcore Technical Liaison Office, "Japanese Telecommunications," unpublished paper, p. 2-1, July 1990).

4. U.S. Department of Commerce, National Telecommunications and Information Administration. 1991. The NTIA Infrastructure Report: Telecommunications in the Age of Information, NTIA Special Publication 91-26, U.S. Government Printing Office, Washington, D.C., October.

5. Gilder, 1991, "Into the Telecosm," p. 156.

6. Arthur D. Little, Inc., 1991, Can Telecommunications Help Solve America's Transportation Problems?, Arthur D. Little, Inc., Cambridge, Mass., May, as reported on in: Passel, Peter, 1991. "The Faxes Are Coming," New York Times, April 10, p. D2.

7. The NTIA infrastructure report referenced above (see note 4) argues for increased competition in relevant markets as a stimulus to the development of infrastructure, and it also advances an updated concept of universal service, which it calls Advanced Universal Service Access (Advanced USA). The report recommends:

8. Lopez, Julie Amparano, and Mary Lu Carnevale. 1990. ''Glassed Houses: Fiber Optics Promises Revolution of Sort If the Sharks Don't Bite," Wall Street Journal, July 10, pp. A1 and A10. The NTIA infrastructure report referenced in note 4 comments on the lower rate of ISDN implementation in the United States as compared to several other countries (pp. 183–187), and it also references efforts to deploy fiber to the home and fiber to the curb (pp. 100–109) as part of the improvements to transmission systems under way.

9. Although the SSA implemented a comprehensive systems modernization plan during the 1980s that brought its information systems back from the brink of failure, a recent assessment concluded that the "SSA can do a great deal more to exploit the benefits of on-line automation" (National Research Council, 1991, Elements of Systems Modernization for the Social Security System, and 1990, Systems Modernization and the Strategic Plans of the Social Security Administration, National Academy Press, Washington, D.C.).

The IRS is itself on the threshold of a major Tax Systems Modernization program that will significantly enhance its ability to exploit the benefits of dealing with information in electronic, as opposed to paper, form. That program is under evaluation by the Computer Science and Telecommunications Board.

10. Kleinrock, Leonard. 1991. "ISDN—The Path to Broadband Networks," Proceedings of the IEEE, Vol. 79, No. 2, February, p. 112.

11. Dorros, Irwin. 1990. "Calling for Cooperation," Bellcore Exchange , November–December, pp. 7–8.

12. The impediments and other issues are similar to those found impeding progress in systems technology generally and notably in the area of software engineering. A discussion of these issues in the context of software engineering can be found in the following report: Computer Science and Technology Board, National Research Council. 1989. Scaling Up: A Research Agenda for Software Engineering, National Academy Press, Washington, D.C.

13. Boehm, B. W. 1981. Software Engineering Economics, Prentice-Hall, Englewood Cliffs, N.J.

14. Brandt, Richard. 1991. "Can the U.S. Stay Ahead in Software?" Business Week, March 11, p. 98.

15. Computer Science and Telecommunications Board, National Research Council. 1991. Computers at Risk: Safe Computing in the Information Age, National Academy Press, Washington, D.C., p. 7.