Part 3
Public Investment in Telecommunications Infrastructure



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 173
The Changing Nature of Telecommunications/Information Infrastructure Part 3 Public Investment in Telecommunications Infrastructure

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure This page in the original is blank.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure Introduction to Part 3 David G. Messerschmitt This session of the workshop discussed different mechanisms for increasing investment in infrastructure through direct or indirect public investment. Areas for investment or subsidy that were specifically considered were basic research, development, service and information providers, and end users. A recent report from the Clinton administration, National Information Infrastructure: Agenda for Action (IITF, 1993), laid the groundwork by suggesting some areas for direct investment or subsidy. It proposed the basic thesis that carefully crafted government action can complement and enhance the efforts of the private sector, and assure the growth of an information infrastructure available to all Americans at reasonable cost. This thesis addresses two phases of evolution of a network infrastructure where free market mechanisms are inadequate, as identified by Eli Noam during Panel 2's discussion: the lack of critical mass early in its evolution (when the marginal utility to users does not justify the cost) and the desire for universal service in the mature phase (when the network is widely but not universally available, disadvantaging a subset of the population). Public investment or its close cousin user subsidy may be needed in the early phases to "jump start" the infrastructure, and cross-subsidy or user subsidy may be required later to extend the infrastructure to all users, particularly disadvantaged people or those in rural areas. Four specific actions related to public investment or action to encourage the national information infrastructure (NII) are identified in the Clinton administration report: Help the private sector develop and demonstrate technologies and applications. This most likely takes the form of government support of research or may take other forms such as tax credits or subsidy of private research. Adjust federal procurement policies to provide incentives for the private sector to contribute to NII development. The question here is whether the government is in a position to choose the winning technologies. Make the vast reservoir of government information available via the NII. Since the government at all levels has traditionally provided library service, to what extent should this extend to becoming an NII information provider? Extend the universal service concept of the telephone network to the NII. Since the NII is likely still in the "critical mass" phase, it may be premature to consider this issue. Nevertheless, there are significant issues inherent in introducing the new NII into the existing telecommunications environment, with its existing cross-subsidy structure and regulatory disincentives to the introduction of new technology.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure Walter Baer gives an enlightening perspective on past government investments in telecommunications, beginning with the construction of the first telegraph system. Starting with the formation of Western Union and later the Bell system, most telecommunications infrastructure was constructed with private funds under the watchful eye of state and federal regulators. However, the government did support research and development of may telecommunications technologies for defense needs, and many of these later became important commercial successes, for example, microwave radio, computing, digital signal processing, satellites, and computer networking. Now whether the government should continue to support technological development, and, if so, how it can best shift from a defense- to a commercial-based investment. Baer describes some ways that government is currently investing in telecommunications and other technologies: Research and development tax credits. A shortcoming of the current approach is that software tax credits are more difficult to obtain and yet are the lifeblood of the telecommunications industry; Direct investment in research and development, much of it subsumed under the High Performance Computing and Communications Initiative; Networks and systems. This includes a direct investment in networking infrastructure in the National Research and Education Network program. It also includes military systems such as the ARPANET and the Global Positioning System that are now making a transition to important commercial uses. It also includes investment in infrastructure for internal government use, such as FTS-2000, that has a substantial impact on commercial technologies because of the size of the procurements; Subsidizing of user networking, which is relatively infrequent but includes support for rural telephone networks through the Rural Electrification Administration telephone loan program; Support of applications, such as telemedicine, which is spread through a number of government agencies and is relatively small in the aggregate; Subsidizing of users, which is also rare but includes support for school and college connection to networks and support of public television stations; Direct funding of agencies that regulate or set standards, such as the Federal Communications Commission, the National Telecommunications and Information Administration, and the National Institute of Standards and Technology. These programs have tremendous leverage over private investments in telecommunications, especially in comparison to the size of government expenditures; and Support for information and databases. The government is the largest creator, collector, user, and disseminator of information. Baer makes the important point that all federal investments in telecommunications infrastructure amount to between $0.5 billion and $2 billion per year (depending on how broadly you define

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure infrastructure), which is tiny in comparison to the roughly $50 billion per year spent in the private sector. Since the government is not in a position to greatly increase these investments, direct government investment will remain tiny in comparison to private investment. Thus, the government will make the greatest impact if it focuses its investments in areas that will maximally leverage private investment. There are three important rationales for government investment where market mechanisms fail: redressing underinvestment in research and development, achieving critical mass in new technologies, and achieving equity in access. Given budgetary limitations, government can have a much greater impact through policies, standards, and regulations than it can through direct investments. The exception is research and experimentation, where government expenditures are a significant factor and are justified by shortcomings in market mechanisms. Baer finishes by asserting that given these factors and the shift of government attention from military to economic and social purposes, the government should shift from the supply to the demand side, focusing more heavily on end-user applications and end-user support, rather than on networking infrastructure. It should also focus more on the diffusion, as opposed to just the development, of new applications. This may include some direct support of end-user groups, such as a voucher system. Baer also urges skepticism in accepting arguments that investment in telecommunications infrastructure will yield compensatory savings in administrative or other costs. Bridger Mitchell discusses direct government construction of network facilities and subsidies to end users. He reiterates the issues relating to the critical mass problem in the early stages of a new network technology and the dissemination or universal service problem that arises as the network grows and matures. Most rationales for government subsidies of networks and users redress one or the other of these problems. Charles Jackson considers government investment from the perspective of telecommunications carriers. He points out that the capital flows and usage charges are huge in relation to any credible government expenditures, and thus the government is unlikely to make any significant difference by direct investment in facilities or subsidies to users. Rather, the government should concentrate on the tremendous leverage it has in regulatory policy, for example, bringing the depreciation time of assets into line with the increasingly rapid obsolescence of the equipment. William Gillis speaks from the perspective of an information supplier. He argues strongly for allowing marketplace mechanisms full rein, and against making prior assumptions about the information marketplace based on historical conceptions of the technology. Government should invest in the timely creation of standards for network interoperability and the establishment of testbed facilities but not the direct creation of services, where there is adequate private-sector activity. An exception is the creation of applications within the scope of government activities such as education. Gillis reiterates that applications never move from zero to a mass market quickly and that discussions of universal service in information services are therefore grossly premature. In the area of information services, he urges the government to make its vast databases available electronically but in the process to also reorganize the data and invest in the development of a uniform nomenclature to make it more accessible. The goal should not be simply to unleash runaway data, but rather to turn the data into easily accessible and useful information. Robert Kahn addresses government investment in research and development. He begins by describing a conceptual framework of a generic network core, a demand-specific layer, and applications built on top of that (similar to the computer central processing unit, operating system, and applications). He emphasizes that applications are what matter to the end user but that the government has a critical role in coordinating and standardizing the core and demand-specific layers. It is extremely important that the core be designed in an "open system" fashion so that market mechanisms can freely define and deploy the applications without interfering with one another. The government should also ensure that there is a level playing field for all commercial participants. Kahn describes in some detail the conception and development of standards for the Internet as a premier example of how relatively small government investments in research and experimental

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure testbeds can have a tremendously leveraged impact on the technology, economy, and society. He also gives a number of examples of companies (such as Sun Microsystems and Silicon Graphics) that have been the result of support by the Advanced Research Projects Agency to universities and another (Cray Computer) that was strongly influenced by government procurement and applications. His basic point is that government investment in basic research and precompetitive technology development can pay great dividends for the economy. Laura Breeden, then of the Federation of American Research Networks (FARNET), speaks from the perspective of a systems integrator who repackages network services for nonprofit organizations. She emphasizes, as did earlier speakers, that the success or failure of the NII will be founded on the applications and the user friendliness and cost of the access devices, not the transport of the network. Also needed is a more active assessment program for determining the value of government investments and helping to guide future investments, rather than simply following the "religion" that the NII will be built and the benefits will be there. Finally, she appeals for greater scrutiny in the areas of massive federal investment, such as the defense and intelligence communities, as opposed to the relatively small scale of investments in research and educational networking. Can those investments be targeted for a greater commercial payoff? The discussion following the panelists' presentation brings out a number of interesting issues: Questioners reiterate the importance of the applications and their value and friendliness to the end user, rather than the networking technology itself. The distinction between private and public investment in the telecommunications infrastructure is perhaps not as great as one might think, since virtually all taxpayers are also regulated ratepayers. Mistaken investments are subsidized by either the taxpayer or the ratepayer, which are one and the same. The NII will be used for accessing information services perhaps as much as for people-to-people communication. The question arises as to what universal service means in that context. What good does it do to get poor people or rural residents connected to the infrastructure if they cannot afford to access the information sources? In addition, an increasing portion of the cost is in access devices (e.g., computers) rather than in network access, and again this is an important obstacle to universal service. Government should consider and assess not only the economic aspects of the NII but also the impact on society. In the past, major new telecommunications infrastructures such as those for the telephone and television have had profound impacts, both positive and negative. REFERENCE Information Infrastructure Task Force (IITF). 1993 National Information Infrastructure: An Agenda for Action. Information Infrastructure Task Force, Washington, D.C., September 15.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure Government Investment in Telecommunications Infrastructure Walter S. Baer INTRODUCTION U.S. government investment in telecommunications infrastructure dates back at least to 1843, when Congress appropriated $30,000 to support construction of Samuel Morse's American Telegraph System between Baltimore and Washington, D.C. (Thompson, 1947, pp. 16–19). In the 150 years since then, the federal government has supported many other advances in telephony, terrestrial broadcasting, satellite communications, and data networking. But unlike investments in highways, mass transit, water supply and treatment, and related categories of physical infrastructure, telecommunications investments have been made largely by private-sector firms under regulations set by federal, state, and local authorities. Consequently, direct government investment in the U.S. telecommunications infrastructure has remained quite small compared to investments by telecommunications common carriers and other private firms. In the past, federal investments in telecommunications have been driven principally by national security considerations and other direct government missions. Defense spending during World War II and throughout the Cold War period spurred the development of microwave and satellite transmission systems, computers for switching and network control, and new network concepts such as packet switching. Today, however, public concerns are shifting from military security toward U.S. economic competitiveness and societal needs in such areas as health care, education, and the delivery of public services. The growing interest in telecommunications and information infrastructure reflects these economic and social policy concerns. Like other infrastructure elements, telecommunications and information networks directly and pervasively support both public- and private-sector activities. Their expanded use is generally believed to increase productivity and output, spur innovation, and lead to significant changes in organizational structure.1 And they are changing rapidly. The policy importance of technological advances in telecommunications and information is twofold. Not only do these advances promise even greater benefits to individuals and organizations who use them, they also break down the regulatory fences that governments have erected to keep industries separate. Technology is rapidly blurring the boundaries between the telephone, cable, broadcasting, and computer industries; between point-to-point communication and mass communication; and between communications and information services. These changes raise anew questions of how public- and private-sector responsibilities for infrastructure investment should be divided. This paper discusses the issues surrounding public-sector investment in telecommunications infrastructure. It basically addresses the following questions: In what ways and for what reasons does the government invest directly in telecommunications infrastructure? Should these investments change in light of the rapidly advancing technology and increasing importance of telecommunications

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure and information? The paper primarily focuses on federal investments in telecommunications infrastructure, recognizing that state and local governments play important roles as well (Mechling, 1993). Examples include state government investments in telephone and data networks, state university investments in local area networks, and local government investments in police, fire, and other communications networks and facilities. This discussion of direct public investment complements the prior workshop session on how government influences private infrastructure investment through legislation, regulation, spectrum allocation and assignment, standards setting, and other ''indirect" means. It should be emphasized that these indirect government activities have considerably more effect on total investment in telecommunications infrastructure than do direct government investments or subsidies. A Note on Definitions Defining the telecommunications infrastructure is tricky, particularly if the intent is to differentiate the "telecommunications infrastructure" from the "communications infrastructure" or the "information infrastructure." The 1991 infrastructure report from the National Telecommunications and Information Administration (NTIA) distinguishes between the "communications infrastructure which, broadly speaking, encompasses all of the facilities and instrumentalities engaged in delivering and disseminating information throughout the nation" and the more narrowly defined "telecommunications facilities, by which we mean facilities that permit point-to-point, two-way transmission of information of the user's choosing" (NTIA, 1991, pp. 13–14). The broader NTIA definition seems akin to the current "national information infrastructure" that includes point-to-point communications, mass media, the U.S. Postal Service and express delivery services, publishers and printers, and the motion picture and video industries. This paper defines telecommunications infrastructure as encompassing all electronic communications, but not print. Unlike the NTIA definition, it includes both point-to-point and mass media such as cable television, since their technologies are swiftly converging and both can be used for the transmission of video, voice, and data services. The definition of infrastructure here also comprises four principal components: (1) the physical infrastructure, including both network and user facilities and their operating software; (2) applications software; (3) research and development on physical infrastructure and software; and (4) human capital investments relating to these other elements. This again is broader than the NTIA definition that includes only the physical infrastructure. FEDERAL INVESTMENT IN TELECOMMUNICATIONS INFRASTRUCTURE The federal government subsidizes or directly invests in telecommunications infrastructure in a variety of ways, ranging from funding research and development (R&D), to building prototypes and operating systems, to supporting users who purchase telecommunications equipment and services in commercial markets. The principal categories of direct government support and subsidy are listed in Box 3 and described below.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure BOX 3 Categories of Federal Investment in Telecommunications Infrastructure Tax Incentives Direct Support of Research, Development, and Demonstrations Internet Other components of the High Performance Computing and Communications Initiative Defense research, development, testing, and evaluation Demonstration and prototype systems Support of Telecommunications Networks and Systems Defense and space systems FTS-2000 and other government systems National Communications System Capital and operating subsidies Development and Support of Applications Support of Users Hardware and software purchases Operating subsidies Training Other Types of Federal Support Support of regulation (e.g., Federal Communications Commission) Support of policymaking (e.g., National Telecommunications and Information Administration) Support of standards (e.g., National Institute of Standards and Technology) Databases and information Tax Incentives Since 1981, the federal government has subsidized technology investments in general through research and experimentation (R&E) tax credits, available to firms whose expenditures on laboratory or experimental R&D are above a base level.2 Most communications equipment manufacturing firms and carriers benefit from R&E tax credits. One controversial issue since these tax credits were first introduced is the extent to which software development qualifies as eligible R&E. Software is an increasingly important part of the telecommunications infrastructure, so that extending the range of software eligible for R&E tax credits would be a stimulus to infrastructure investments. Investment tax credits were eliminated in the 1986 Tax Reform Act. Legislation to reenact them has [previously] been proposed to stimulate overall U.S. investment, but it seems unlikely to pass in the current Congress. Direct Support of Research, Development, and Demonstrations The administration budget for fiscal year 1994 (FY 94) included $72 billion for research and development, of which $30 billion (42 percent) was for nondefense purposes. Most of the

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure nondefense R&D support for telecommunications, as well as $343 million from the Advanced Research Projects Agency (ARPA) of the U.S. Department of Defense, is now subsumed under the $1.1 billion multiagency High Performance Computing and Communications Initiative (HPCCI). The HPCCI includes five principal components:3 High-performance Computing Systems Research for future generations of computing systems System design tools Advanced prototype systems Evaluation of early systems Advanced Software Technology and Algorithms Software support for "Grand Challenges" Software components and tools Computational techniques High-performance computing research centers National Research and Education Network (NREN) Internet Gigabit research and development Basic Research and Human Resources Basic research Research participation and training Infrastructure Education, training, and curriculum Information Infrastructure Technology and Applications (IITA) Information infrastructure services Systems development and support environment Intelligent interfaces "National Challenges" The NREN and the new IITA program are the two components most clearly directed toward advancing the nation's telecommunications infrastructure. However, the entire HPCCI can well be considered to be a federal investment in telecommunications and information infrastructure. Beyond its contributions to the HPCCI, the Department of Defense invests heavily in advanced communications research, development, test, and evaluation (RDT&E). The FY 94 line-item budget for defense RDT&E for "intelligence and communications" totaled more than $5 billion. Although it is difficult to estimate how much of this directly related to the national telecommunications infrastructure, defense requirements have in the past often led to technology developments with commercial applications. Examples of such national security programs include much of the early work on digital signal processing, R&D on the ARPANET, and research by the National Security Agency on encryption in commercial telecommunications networks. The federal government has occasionally supported prototypes and demonstrations of telecommunications systems intended for widespread commercial use, of which the Morse telegraph system is the earliest example. In the 1970s, the National Aeronautics and Space Administration (NASA) sponsored the ATS-6 satellite program that successfully demonstrated direct broadcasting of television and other communications services from space. The ATS-6 demonstration convinced a

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure number of developing countries to invest in communications satellite systems, thus benefiting U.S. suppliers of satellites and ground stations (Cohen and Noll, 1991). It did not, however, directly influence the pace of investment in satellites for the U.S. telecommunications infrastructure. In September 1993 NASA launched the Advanced Communications Technology Satellite (ACTS), which is a direct descendent of the earlier ATS program. As its name implies, ACTS is intended to demonstrate advanced technologies—principally high-data-rate switching and processing onboard the satellite. NASA project managers say that ACTS has directly influenced commercial infrastructure investments such as Motorola's Iridium system, which will provide satellite links to wireless communications users (NASA, 1993, pp. 1–2). Critics claim, however, that commercial technology had bypassed the $600 million ACTS program well before its launch (Broad, 1993). Support of Telecommunications Networks and Systems The Department of Defense and other national security agencies have traditionally built and operated their own telecommunications networks, ranging from secure military communications links to the ARPANET, which began as a low-cost computer-to-computer network, to the Global Positioning System (GPS), which sends satellite signals that give accurate location information worldwide. Some defense networks such as the ARPANET and the GPS have evolved to become part of the telecommunications infrastructure, but most remain separate systems dedicated solely to defense missions. A few civilian agencies such as NASA build and operate extensive telecommunications networks of their own, but most other federal agency communications are coordinated by the General Services Administration (GSA). Since 1987, GSA has maintained an Information Technology Fund to procure telecommunications facilities and services for federal agencies. Its largest contract is known as FTS-2000, providing intercity voice, data, and video services. Managed by AT&T and Sprint, FTS-2000 is a software-controlled "virtual network" whose traffic is physically commingled with nonfederal communications. Whether or not the primarily leased services under FTS-2000 constitute federal "investment" in a technical sense, they certainly support investment in the telecommunications infrastructure by private carriers. Through the Information Technology Fund, GSA also procures local telecommunication services and supports information security and emergency management programs. In addition, the federal government maintains a National Communications System Office responsible for providing critical communications needs during emergencies (NCSO, 1993). The history of the Internet illustrates how federal support of telecommunications networks can change as the systems evolve (Hart et al., 1992). The original concept of distributed computer/communication links was based on research in the early 1960s supported by the U.S. Air Force (Baran, 1964).4 In 1969 ARPA built the ARPANET, the first computer-to-computer network based on this "packet-switching" concept. The ARPANET was a fully government-funded network linking the Department of Defense and its contractors. By the early 1980s the ARPANET's success led to proposals to develop similar networks for nondefense uses. The National Science Foundation thus funded its own NSFNET to link its supercomputer sites and other major U.S. centers of computing research. Other regional and local networks quickly developed, linked by the NSFNET and other "backbone" networks and connected to many sites and networks around the world. Today, the Internet comprises tens of thousands of interconnected, interoperable computer networks. Federal direct support represents only about 10 percent of total Internet costs in the United States (Hart et al., 1992, p. 686).5 As of July 1993, more than half of the 46,782 registered networks connected to the Internet were commercial.6 Whether federal agencies should continue to provide network subsidies as the Internet becomes part of the national telecommunications infrastructure has emerged as a key policy issue.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure was focused on the needs of the education and research communities initially, and federal support allowed them to develop new services and develop the market for the Internet by reducing risk, guaranteeing a certain level of demand, and creating a critical mass of users. In late 1993, as we contemplated the development of the national information infrastructure and how the government's investment strategies should change, it became time to ask, "When do you move federal investment from the supply side to the demand side?" (As many other members of this panel have said, it is also important to look at all of the tools at government's disposal, such as tax policy and regulation, since these may be more effective and influential in many situations.) It is doubtful that if Washington gave a $100 tax rebate to every teacher in America for "networking" (a $100 million investment), the result would be a national research and education network. There is a social organizing principle at work here, which argues for targeted spending aimed at leveraging certain activities within the target community, whether that is K-12 education, libraries, health care, or economic development. To be effective, federal investment policy will have to be designed to meet the needs of specific user communities, at specific points in time. Generalizations about information infrastructure are becoming rapidly more hollow as new technologies, service providers, and access devices sweep on to the scene. Lessons from the past will be useful only to the extent that they are carefully drawn from particular historical circumstances. Policymakers will need to practice humility. They will need to be painfully honest about what they know and what they don't. Data, and the recommendations that follow from them, will have to be looked at with a cold eye. A lot of what we heard at this workshop was amazingly vague, imprecise, or not well supported. Further, we need to maintain a sense of scale in policy discussions. The U.S. telecommunications industry spent approximately $200 billion in 1992. Comparatively, as mentioned in this workshop today, defense research, development, testing, and evaluation for intelligence and communications accounted for $5.3 billion, the High Performance Computing and Communications Initiative (HPCCI) accounted for $1 billion, and the National Research and Education Network component of the HPCCI accounted for about $100 million. Typical research university computing cost $50 million, the National Science Foundation's Networking and Communications Research and Infrastructure Division budget was $35 million, and the NSFNET backbone was $12 million. What can the federal government do effectively? Is there a role for government investment? I believe that government has an obligation to promote equity, in the first place. Further, public support at the right time and in the right place can offer: Risk reduction, in high-risk but potentially promising areas; Proof of concept (when private capital is not available, because of the risk or because the benefits to be obtained are not likely to attract private investment); and Support for network externalities (such as the development of a critical mass of users) and for public goods. There are a number of ways to achieve these goals, including support for: Applications development; User connection to the network or information service; Development of telecommunications networks and systems; Direct R&D; and Incentives (e.g., tax policy).

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure I concur with Walter Baer's excellent test of investment policy as presented earlier at this workshop: Involve industry and users in design and deployment; Require cost sharing; Have an explicit plan for handing off technology to the private sector; Focus on meeting technical and economic objectives; Shelter from capture by political constituencies; and Avoid rigid schedules and time constraints. I would further recommend that formal evaluation and dissemination be required components of all federal technology and infrastructure investments. To summarize, I urge that policymakers: Examine their data critically and carefully. Data must be accurate, and conclusions that follow must be relevant to specific times and circumstances; Admit their assumptions up front, and question them continually; and Maintain a sense of scale. Data presented at this workshop demonstrate that some of the areas that have received the most attention from policymakers actually loom small within the overall dimensions of federal activities. Pay attention to the things that matter. REFERENCE Mandelbaum, Richard, and Paulette A. Mandelbaum. 1992. "The Strategic Future of the Mid-level Networks," pp. 59–118 in Building Information Infrastructure. Harvard University Press, Cambridge, Mass.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure Discussion JOEL ENGEL: I would like to address a couple of things. The first is the whole issue of critical mass, which so many people have spoken about. Surely, there is a critical mass issue. But that doesn't require subsidies. We have many examples, even from very recent history. For example, the very first purchaser of a VCR [video cassette recorder] found it not very useful if there were no cassettes to play on it, and it would not be very good business for people to make cassettes if there weren't very many VCRs. In the same way, without software the utility of PCs [personal computers] would be considerably less than it is today, but it was not very desirable for people to develop software if there wasn't a large mass of PCs and PC owners to sell it to. And yet, those are two examples of exploding industries that didn't require any kind of subsidy to get started, but were started by what we used to call investment. Wally Baer gave the example that even for the telephone industry, the entire telephone network was built on private investment. The government did invest. It invested in telegraphy, which is the other example of the ability to choose winners. Bob Lucky gave the figure of about $25 billion as being required to convert the telephone network from copper to fiber cabling. But we ought to put it in perspective. As Chuck Jackson said, just the LECs [local exchange carriers] alone spend about $20 million every year to keep their networks modern. In the last few years about a third of that has been in local distributions, but that is because all of us have been replacing our analog switches with digital switches. That emphasis can shift. In addition, the front page of the business section of yesterday's USA TODAY shows that the free cash flow of the LECs ranges from $4 billion to $6.5 billion, which means that even in addition to this shift, if there were good business reasons, those investments could as much as double. So, as Wally Baer pointed out, while we sit in this room and debate whether or not there is a critical mass issue, the private investor is going to inconsiderately knock the legs out from under the debate. The other point that I want to make with much fervor has to do with the view that one of the necessities is to make the upgraded infrastructure available to all Americans. I don't think that is the critical issue. The critical issue is to make it interesting and attractive to all Americans. I really want to reinforce the point that Bill Gillis made about the importance of applications and content. Earlier today, both Bruce Alberts and George Turin told about having to either hide or lock up their TV sets because they didn't like the content that their children were getting out of it. They didn't say, but from my experience as a parent, I am willing to bet that they didn't have to lock up their encyclopedias. I think that the real challenge is understanding what, when we build the infrastructure, is the content that will make them come. DAVID MESSERSCHMITT: I think that particularly on the first point—that market mechanisms will take care of the critical mass problem—there may be some comment.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure CHARLES JACKSON: Just a historical observation. I believe that the first widely known economics paper on critical mass issues in networks, network externalities, is by a colleague of mine named Jeff Rohlfs [formerly] of Bell Laboratories. He looked at the problem by looking at the videophone experience in the late 1960s, when the old Bell system rolled out a product for which they charged $60 a month independent of usage. His analysis was that if the charges had been based on usage so that people had the product on the desk but didn't have to pay for it if they didn't use it, the network might have grown a lot faster. Economists have tools to analyze these issues. But I think that the insights that we get from network externality theory do give us clues about how to price new services, and maybe some ideas for promotional activities by the government. I think the point in the long run—which complements Bridger Mitchell's point that if you try to subsidize something and then it never quite makes it to the critical mass point, then you still have all these people getting the subsidies—is that you can lock yourself into a very unfortunate and wasteful situation possibly, and we need to be sensitive to that. But nevertheless, I think that theory can give a lot of useful insights. MESSRSCHMITT: Would those pricing mechanisms be consistent with making a profit during this critical mass period? JACKSON: No. What you need at the beginning is a promotional price to get people into the network, and then you profit from it later. And clearly you can imagine industry structures where that second stage isn't possible. WALTER BAER: I don't think you are hearing any wild enthusiasm from this panel for large government investments in this area. What you are hearing, I think, is that there may be government roles on the R&D side in dealing with some of the network critical mass issues. The two examples that you gave are interesting. The VCRs and the PCs really are kind of stand-alone products, at least initially. They didn't require networking to be sold initially. And there may be some difference when you get this network externality effect. The other point is that the VCR actually did profit from a fair amount of government-supported R&D 10 or even 20 years before; much of it was, again, defense oriented at the beginning. LAURA BREEDEN: I agree with the point, though, that the applications that are going to be delivered over the information infrastructure are critical. I heard Larry Irving, who at one point was very involved in the reevaluation of cable before he went to the Department of Commerce, say that the two biggest sellers on pay-for television are tractor pulls and soft pornography. I believe that the government may have an interest in seeing that other kinds of programming are available that perhaps are not driven by the market, because the market either isn't there or is so immature that no single private investor or group of investors can make a profit on those applications. Having said that, I don't think it is fair for the private sector to say that government should fund all the high-risk application development and that the private sector then will reap the benefit, the profits. One of the things the questioner said was, "let's make it attractive," and that is the hard part. If there is one thing that U.S. industry is good at, it is making stuff attractive. It is creating demand. So, I think there is an issue of balance here and it is not a black-or-white decision between no government investment and socialized networking. LUCY RICHARDS: Since I come from an R&D committee on Capitol Hill, I want to make a couple of comments on the issue of government R&D. First, I appreciate the comments on the National Information Infrastructure Act of 1993. That was our committee's bill and is basically a refinement of Public Law 102–194, which funds a lot of the high-performance computing applications. As that bill went through our committee, we tried very hard to refine it, to try to make clear that the government was not trying to compete with the private sector in the development of services or networks. That bill also makes significant changes to the NREN programs, with one of the goals being to make clear that the NREN is not trying to duplicate what private industry can provide.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure A couple of other comments on funding. One concerns the issue of manufacturing, which we have addressed because we found it very relevant. It is contained in H.R. 820 as a separate provision. It is not a high-performance computing application. It is in S. 4, though, as a high-performance computing application, which is development of applications using information technologies to assist manufacturing. There is, in H.R. 820, the development of a manufacturing outreach program that will have as part of it an advanced communications network to help link small- and medium-sized manufacturers and give them assistance, because we found that even though there are manufacturing centers around the country, small business owners can't take the time to travel to them to get assistance. So we are going to try to link them up electronically, in a kind of extension network…. One last comment on financing. In H.R. 820 we have a Title III, which essentially provides for the government to set up some venture capital mechanisms, funneling money into other organizations that will hopefully provide venture capital financing for the development of critical technologies. There are a lot of communication technologies on the critical technologies list. The administration hadn't accepted that. They agreed to make it a study. But at the same time, when you look at the National Performance Review, Sally Katzen, when she testified before us, was talking about a program to finance the development of information technology applications for government purposes through a venture capital fund. I think it was 1 percent of the agency's operating budget. Then you have venture capital provisions in the seed act that was passed a couple of years ago, which has the government in the position of giving government money to finance venture capital, start-up funds in Eastern Europe but not for American companies doing the same thing in the United States. What we have heard is that the Treasury Department has decided to set up an interagency group to look at the whole issue of government financing for new technologies. BRIAN KAHIN: I think Joel Engel's example of the VCR was a bad one, because the VCR business took off on the basis of off-air taping, rather than the existence of cassettes. And I think it is a problem with the PC analogy, too. WILLIAM GILLIS: Maybe I can help just a bit. I introduced the VHS-brand VCR in August of 1977, and in fact, it did take off immediately. There was a predecessor, the Betamax, the one-hour machine that was out in 1975. We introduced the VHS in 1977 as a time-shift mechanism and to give the ability to record off the air. From the day that we first introduced it, I went into backlog for the next year and a half because of the incredible demand. One story that I will share with you highlights the point. We were bold enough to put my name—it was an error—in one of the ads that we ran. I got a call from a lady who said, "Are you Mr. Gillis? Are you the one that is responsible for this VHS brand VCR?" I said, "Yes." She said, "Well, I called to say thank you.'' For the first time in 2 years her husband had taken her out to dinner on a Monday night because he was able to record Monday night football. Then she said, "I have a problem. Last Monday night's game went into overtime and he hasn't watched it all yet and we are out of blank tapes. Could you please find me a way to get blank tapes?" And that was only one of many such calls. Clearly, the overwhelming evidence—and there is certainly data available—is that off-air recording was the driver; you can record when you are not at home or record when a competing program is running. That information is very definitive. BREEDEN: What was the cost of that device when it was introduced in 1977? GILLIS: We introduced the first one at $1,000 suggested retail. We sold it to distributors for $750. We made money the first year. MICHAEL ROBERTS: I want to make a brief comment on the issue of winners and losers and the federal role in that, which gets everybody's blood moving. I was surprised, with all of the economists in the workshop, that nobody has mentioned Christopher Ferrell. One of his observation

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure is that one of the great advantages of capitalism comes through innovations and asset reallocation. It harnesses a lot of creative energy and improves society. But intrinsic to that is the notion of whether an asset reallocation is going to destroy industries. When diesel engines did away with steam engines, or Henry Ford did away with horses, thousands of jobs were destroyed, but that wasn't considered a big government policy issue. Since World War II, the government has become the majority force in technological innovations. So, the government is now linked to the destruction as well as the creation of a tremendous number of jobs, which means that there isn't any way for us to get the winners and losers out of this policy space. I think that the challenge that we have is evolving a policy space around the NII [national information infrastructure], which we really haven't done—certainly this day and a half has shown how primitive the policy space is around the NII. We have to manage that issue, not wish hopefully that it will go away. KAHIN: There is also the example, on the other side, of the Internet—network externalities, the government leveraging with an extremely small amount of money what has turned out to be an outstanding and increasingly commercial success. The general information infrastructure development problem has been well phrased. There are many arguments for government investment. A couple Wally Baer mentioned that haven't been pointed out are the information cost and the transaction costs, which is why the government investment of the producer's subsidy into the backbone has made sense at the beginning. The basic problem is the information management problem and the risk of the government interfering in private investment. If the government doesn't have the information flowing to it in a way that enables it to refrain from stepping on private investment, it is not going to work. And that is a problem in the applications area, too, that we really haven't faced yet. I was struck by Bill Gillis's comments that he has been relying on patents in this piece of the infrastructure he is building. This is a situation where the government is directly involved in issuing patents, and it is an information management problem. The government is doing an absolutely crummy job of handling patents in the software area. It is a disaster. It is taking 32 months to process the applications, and the quality control is absolutely abysmal. At the same time, what you have witnessed in industry in general is a moving away from proprietary positions because the users won't accept it. So, you are building a proprietary system, you are placing your bets on a proprietary infrastructure system, and you are taking a big gamble that there won't be some kind of standard that passes you by. GILLIS: On that, I think you misunderstood. While we do hold parents, we are not necessarily basing our system on patents as its protections. And you will find that the customers, who number among the household names for the products and services, would differ. If I left the impression that that was our sole safeguard, I want to point out that what we are going to deliver is the highest-quality service that is available. It is not an issue of proprietary software or not that is really driving our market, because we are taking the information of our various customers and we are delivering it back to them in a form that they and their customers find useful. CAROL HENDERSON: I want to thank the National Research Council for including various user communities in this workshop, such as health care and education and libraries. To misquote a former FCC chair, this particular infrastructure is not just about selling toasters. Unlike, say, the agriculture or the transportation infrastructure, it really is the infrastructure that is related to what makes us human, the ability to communicate. And it certainly has the potential to be in the future the main way that humans communicate beyond just shouting distance to each other. As such, it will have enormous and probably unpredictable effects on society at large. In the discussions on how we go forward, what the government role is, and how we rearrange the regulatory landscape, I think that the effects on society are something that must be considered, if it is not just economic stimulus alone. Concerning some of the government roles we

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure talked about, there seems to be some general consensus on what might be useful, such as stimulating certain kinds of applications and fostering at least start-up activities for certain communities. Those roles are very useful, and I support them. But in a way, I take libraries, which I represent, as an example. Giving a library a little bit of money to get started in doing something new in the telecommunications or computer networking area is very useful. It requires that institution to come up with some matching money and stimulates a lot of additional activity. But it doesn't necessarily solve the systemic problem of making sure that that publicly funded institution is able to get from the infrastructure what it needs to continue its mission in the future. And some of what, say, schools and libraries will need is similar to what other communities need in terms of ubiquity and interoperability and a fair amount of capacity and so on. But there may be other things needed that don't necessarily coincide with the mass market approach—pricing mechanisms, for instance, such as the ability to predict what that cost is going to be and to have that cost not be too usage sensitive. Otherwise use in the education, research, and library areas will be hindered, which actually is harmful to society. These are points that we need to pay attention to in the basic regulatory structure. BAER: I think you raise very good points, and you bring us back to the question of pricing and appropriate regulation, which is very real. I also relate to your early comment about selling toasters because, interestingly, we are dealing with different cultures in the various media, in terms of how we are generating universal access. Obviously, in the broadcast media, the way we get near-universal access to low prices is to use advertising in support of content. It is interesting that the culture of the Internet so far has not only avoided that kind of service, but has also actively rejected it in many cases. I think it is an issue, as we move ahead and look toward larger mass use of the Internet-related networks, as to whether we are going to allow selling toasters to be a part of the system. ALFRED AHO: I would like to make two points that I don't think have had adequate attention during this workshop. One is the importance of software in the information infrastructure. A number of industries today, say, the telecommunications industry, run on hundreds of millions of lines of code in operation. To develop 100 million lines of code could cost anywhere from $10 to $100 per line of code. We are talking about investments of tens of billions of dollars to create the software to run the infrastructure. I just want to make sure that people understand the importance of software. Yet, there is another consequence to this, and that is the embedded base of software that is already there. And this embedded base may be more of a bottleneck than many people realize. Companies and universities just can't afford to get rid of—or the Internet, in particular can't get rid of—its embedded base of software overnight. Bill Gillis mentioned the importance of standards and interoperability. I would like to add a third point, and that is evolvability of the infrastructure. The technology is changing very quickly, and the applications are also changing very quickly. Whatever kind of infrastructure we put in place should be an evolvable one, so that 20 years from now we won't be stuck with a narrow-gauge railway system that we can't get out of. We have seen somewhat similar effects already with the current infrastructure that is in place. The second observation that I want to make with respect to direct investments (and in particular, whether this is something for the government) concerns investment in people, people who understand not only how to create the new infrastructure, and the evolving infrastructure, but also how to use the infrastructure. Vice President Gore talks about reinventing government, and we have had a number of excellent presentation about how, by changing business processes or the process we use in various segments of industry and other human endeavors in health and education, we can do business and restructure the manner in which we interact with one another and conduct business in those arenas.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure I think we are just now beginning to see, in a very early stage, the profound impact of this kind of business prospect—restructuring—that is going to take place in almost every area of human endeavor because of this infrastructure. What we need to do is to train people to be able to take advantage of this. ROBERT KAHN: I couldn't agree more about the importance of software. Between the physical infrastructures and the end-user applications is the so-called middleware, which is largely services based on software. It is, in some sense, the network analog of the operating system, plus some more things, that is all software-based. And obviously, investing in people is what you ultimately end up doing with your investments. I agree with Al Aho completely. I would like to comment about that $5 billion number that Laura threw out, just for clarification. Basically, as I understood that number, it goes for a number of things that are unique to defense. Defense has a centralized capability for communications and also a distributed part that the Army, the Navy, and the Air Force separately administer. They have their command and control messaging systems, which are part of the command structure of the military, for getting orders out and controlling our forces. They have a command and control backbone that they maintain. It is secured and protected against a variety of threats. They worry about security technology a lot. They worry about systems that operate on the ground, in a field headquarters of the Army, and on ships, planes, and a variety of space-based systems like MilStar and the GPS [global positioning system], and they also provide support for a lot of other organizations. The R&D involved in those systems, often unique, one-of-a-kind military systems, is large. That is what the R&D budget goes for. The actual maintenance and operation of that stuff can be much larger. DALE HATFIELD: While we are congratulating ourselves on the success of the Internet, I think we should remember something that the FCC [Federal Communications Commission] called the enhanced service exemption, and also the special access exemption from the payment of access charges, which means, essentially, that when you make a circuit-switched call, you pay a subsidy, and when you make a packet-switched call, you do not. Obviously the packet-switching community has benefited greatly by that exemption. I think we have to be very careful when we talk about success to recognize that we are not calculating all the costs. I think that should be considered. MESSERSCHMITT: Can you elaborate on that a little bit more, Dale? HATFIELD: Yes. Let me give you an example. My elderly parents live in Ohio and I call them on a regular long-distance network. About half the cost of the call is split between the two carriers on either end. And a fraction of that is a subsidy that goes to maintain universal service. When I call on the packet node in Denver and I sit there for two or three hours, it is a flat-rate call and I am paying no common carrier line charge on the originating side of that or on the terminating side. So, essentially, when I make that packet-switched call, I am not paying the subsidies that I do when I make a circuit-switched call. And that is one of the reasons that a fax call is free on the Internet and yet costs X dollars on the circuit-switched network, because of the payment for access. When I call, an enhanced service exemption of the FCC says I can make a circuit-switched call to that packet network and I am not paying the access charge. It is treated as a local call, even though it is terminating in Columbus, Ohio. JACKSON: When the Commission proposed removing the enhanced service exemption, it became quite contentious. HATFIELD: And we were inundated with computer-generated mail that went to Congress that put political pressure on the Commission to maintain the exemption. JACKSON: Never pick a fight with an interest group whose hobby is word processing! HATFIELD: Let me make one other distinction that I think we are missing here. I think we all agree that most taxpayers are also telephone subscribers, that those two communities are roughly the same.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure There seems to be a consensus against government investment. But then, I hear Chuck Jackson saying that if we just speeded up depreciation a little bit …, and I hear Joel Engel saying that if we have all these billions of dollars of free cash flow, if you just make it a little bit more attractive, we will make this investment. But if you are making an investment as a utility and there are constitutional provisions against confiscation and so forth, if this turns out to be a bad investment—in other words, if the investment doesn't work out—then you and I as ratepayers are going to be stuck with paying for that investment through higher rates. I think some of the distinction here between private investment and public investment is a little bit wrong, because ultimately, as ratepayers, we are going to be responsible for that investment at some point anyway, if it is made under the traditional sort of public utility type of investment. So my conclusion is, let's not kid ourselves. There is not so much difference between public investment and private investment as one might think, from the public utility standpoint. MESSERSCHMITT: Thank you, Dale. Any comments on that? JACKSON: Yes, I want to respond, because there are some differences. One is that, although the budget process is difficult and contentious, once you vote to make the investment, you can go ahead. With private-sector investment and even the regulated utilities, if their management decides that they aren't going to get that money back, then they aren't going to make that investment. We haven't seen that kind of drawing up of investment in telecommunications yet, but I understand that with electric utilities now, about half to two-thirds of the generating capacity being built in this country is what are called qualified facilities rather than being built by the utilities. The combination of prudence reviews, disallowances, nuclear plants, the whole nuclear fiasco, has dried up utility investment in power-generating facilities. If a telephone company chooses not to invest in new infrastructure, it is very hard to order it to do more. It is possible, but it is a tough fight. HATFIELD: On the other side of the coin, if they go ahead and do it and it turns out to be a bad investment, we are not going to let telephone companies go out of business. JACKSON: Well, we might let them go out of business at some time in the future. We will have to wait and see. We have let long-distance carriers go out of business when they imprudently invested in obsolete satellite facilities, things like that. But if it is the only phone company in town, then I think society will be very reluctant to let it go out of business, although we have had some electric utilities lose a huge amount of money in nuclear power plants. I think that nuclear power plants lost something like 10 percent of their equity in the late 1980s from failed nuclear power plants. I think that some people have put forward a variety of schemes, like price caps, for moving away from traditional utility regulations so that you can free up this investment and not face the specter that, if the investment fails, then the residual monopoly ratepayers are going to bear all the costs. But those are very contentious issues, and a lot of people who support reform are quite unwilling to say okay, we will let it go, we will cap the prices, and if the returns run up to 30 percent on equity because of some great new efficiency, we will let them keep it. JOHN RICHARDSON: I believe one of the most important points made by the panel, through Laura Breeden's remarks, was the need for an estimate of the benefits of the NII. I have felt this for a long time and I am very glad to hear this come out. I do remember one of Bob Lucky's popular columns in IEEE Spectrum a few months ago, perhaps a year ago, discussing the NREN. I believe his point of view in that column was build it and they will come. Perhaps much of the NII is based on this same sort of religion, but for firm policy development, I do think that we need some sort of impetus. We need to begin that task.

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure My question for the panel is, Do you have some ideas or some suggestions of how to approach this estimating of the future benefits? BAER: That is obviously a hard topic to quantify at this point, because of both the long-term nature of the investment and the problem of confounding a number of government objectives—the equity objective along with any kind of efficiency ones. One very crude measure of the effectiveness of government programs past the R&D stage is to require matching funds from private industry. How much is the private sector willing to invest to match a dollar of government money? That is not the only measure, but it seems to me it is a reasonable one to place on a whole variety of proposed government investments beyond basic R&D. And on that measure, so far, it looks like the NREN would have come out pretty well. RICHARDSON: You are suggesting that we feel our way. And I am asking for something more than that. BREEDEN: I am not an economist, so I am not putting my professional reputation at stake here and I can say anything I want. Having prefaced my remarks with that comment, I would say that if you are going to make an investment decision as a nation, you want to try to evaluate the long-term or medium-term benefits of that decision. Now, if this information infrastructure succeeds, presumably there are going to be more devices sold, which means healthier industries in computing and personal communications systems and so forth. It means that more transport is going to be sold, which I think is why you see lots of phone companies doing deals with lots of entertainment and information service providers. I think you can probably make a case—and again, I am not an economist—that a lot of spending decisions in the private sector are going to be stimulated by the development of an infrastructure that we can all use. If we look at highways and cars, and gasoline and fast food, there is maybe a similar relationship there. And I think the exercise is worth doing, because, you know, we, as a nation, have some very hard choices to make. Are we going to increase our national debt indefinitely, or are we going to do something different with our capital and our country? KAHN: It is clearly an interesting question to pose, and it would be very nice if we could get some answers to it. You could even broaden the question to ask not only what the benefits are, but also what the return on investment is, and you could phrase that in various different terms. The thing that I would like to caution us on is not holding our breath waiting for an answer to those questions, because I am not sure they are going to be very easily forthcoming. If you were to try to ask that same question in other areas where, I think, people have naturally just bought into the concept, you would have an equal amount of difficulty. For example, if you asked what the benefits of personal computers are in the workplace, I can make a strong argument that they are negative as well as positive, that you are spending too much money and it takes too much time and it has encumbered you. If you were to ask about the value of education, which we spend a lot of money on, you would have a tough time quantifying the benefit of any one year of the educational process, or the whole thing, or even what society would be like if you had more or less of it or a different kind. Likewise, I think you have much the same problem with health care, despite the fact that everybody knows the need for it and wants it. Estimating the benefits of it would be very hard to do. I think you would have an even worse time trying to explain what the benefits were of the economy in general, even though I think you could make a better stab at that than perhaps all the others. So I just wouldn't hold my breath waiting for the answer, despite the fact that I wouldn't mind seeing something that tries to describe the answer. RICHARDSON: I appreciate your point of view and I agree with it. My point of view is to encourage somebody to try to get quantitative, even though it is a feeble first step. BRIDGER MITCHELL: I think Laura steers us in an interesting direction of trying to be specific. But trying to answer the broad question about the benefits of the NII is almost doomed to

OCR for page 173
The Changing Nature of Telecommunications/Information Infrastructure fail from excessive generality. If we can focus on specific programs, specific kinds of initiatives, there is some chance of advancement. For example, we have a particular quantitation of the way federal action has held up gains in consumer value in the cellular area where, for nearly 10 years, it delayed the introduction of a technology that was clearly there. We have market evaluation processes for measuring that sort of thing. But the market may not send adequate signals to evaluate the "public good" aspects of education and consumption of raw information. CLIFFORD LYNCH: In listening to this, I have felt that an aspect is missing in some of the discussions. I want to try to quickly outline it and then solicit some reaction. I think that when we have been talking about this information infrastructure and looking at some of the economic models that might provide insight, we tend to be thinking about a communications network. And we think about universal service objectives in terms of getting everybody connected. I have some reservations about whether the primary use of this new network world is going to be person-to-person communication for the vast majority of users. Certainly, things like the telephone work really effectively for a lot of people right now. And things like the picture phone haven't been roaring successes. I suspect that a lot of the use of this system is going to be access to various kinds of information services. If you think of that as accounting for a lot of the use of an information infrastructure, a bunch of questions come up. What does universal service mean in that sense? You have connected people to something, but they can't afford to use most of the services on it, because they cost many many times what the connection costs. As with the phone, just because we give them universal service doesn't mean that we give everybody a blank check for 900 numbers. Another point. It seems to me that another area of potential government investment choices to be considered is placing information services on your network, including, as you mentioned, some of the enormous stores of government information that could be organized and made accessible, either at very low cost or perhaps, in some cases, for free if there are policy or legislative mandates to do so, for access by the citizenry. It seems to me that that is a whole other dimension of potential government investment that could move along the creation of the infrastructure by giving people reasons to want to make use of it. BREEDEN: I think Cliff is right. It is incorrect to focus only on the communications aspect of this and to assume that that means people communicating directly with other individuals. One of the other panelists made a point about the need to standardize presentation of data. Cliff has done a lot of work in something called Z39.50, which is an information standard. I think those areas are going to be tremendously important. I tend to think that the equity issues, if we are going to avoid having a society of information "haves" and "have nots," are going to be solved in the public libraries. MESSERSCHMITT: Laura made a very important point from my perspective—that is, increasingly, the cost to users in terms of equipment and services is often for equipment that they buy themselves, rather than for network transport services, which are rapidly being driven toward zero. Bob Lucky made this point in his talk, that Internet was so cheap in comparison to the telephone. So I think that the issues with respect to customer-owned equipment that is necessary are quite different, probably, from the issues having to do with the communication infrastructure and yet may ultimately be more important in terms of issues like universal service.