Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
TECHNOLOGICAL TRENDS AND IMPLICATIONS* To appreciate why computer-based information technology presents unusual opportunities and challenges affecting wide sectors of American society, one needs to appreciate the capabilities of the computer, both as an information- processing device and as a facilitator of modern-day com- munications. The appreciation of those capabilities can, in turn, be helped by delineating carefully a handful of simple but often only hazily grasped concepts, one of the most important of which is the concept of information representstion. Consider information as an abstract entity that in some way reveals knowledge about the environment in which each of us livesâthe planet Earth, the solar system, the cos- mos, a community, an institution, or a series of events. To convey such informationâi.e., to pass it from one per- son to another, or from one institution to anotherâit has to be represented in some form. For example, one item of information in a community is each inhabitant's name. An individual's name may be represented as a written signa- ture, as a sequence of printed letters, as a series of spoken sounds, as a sequence of Morse code characters, or in surrogate form as a sequence of numbers on (say) a credit card, a mailing label, or a personal check. Each of these forms of representation fulfills the same functionâthat of conveying a given item of informationâ and each can be converted into the other without altering the item's basic information content. *This chapter is included mainly for the lay reader. Thus, readers with a strong background in communications and com- puting may want to skip at least the first 7 or 8 pages. 12
13 Broadly speaking, there are two categories of ways in which information of any kind can be represented. One is called digital; the other analog. Digital representation is distinguished by its sequences of discrete symbols, usually letters or numbers, that have meaning when organ- ized according to specified rules. English words, sen- tences, and paragraphs are a good example. The symbols they employ are the 26 letters of the Latin alphabet, plus a variety of special symbols called punctuation, which, when organized according to the rules of grammar and syn- tax, make it possible to convey information of almost any kind. Analog representation, in contrast, relies not on the ordered sequencing of symbols, but rather on representation in terms of some continuously varying physical entity or attribute. The automobile speedometer, for example, shows a vehicle's rate of forward motion by the varying position of a needle mounted on a moving shaft. Change in the needle's position, while typically read against a numer- ical scale, does not require a scale to be comprehensible. Other commonplace examples of information conveying de- vices that rely on analog representation include the cas- sette recorder, which represents sound as continuously varying states of magnetism on a tape, and the analog por- tions of the telephone network, which use a continuously varying electrical signal to represent human speech. That there are these two different categories of ways to represent information carries important practical im- plications. Within either category, for instance, there can be a variety of representational forms, any one of which may be interchangeable with others. Perhaps the best illustration of this in the digital category is the use of numerical symbols in lieu of letters to represent information that are not inherently numericalâe.g., let- ters, words, and sentences. By simply numbering the let- ters of the alphabet 1 through 26, the information conveyed by any letter can be made independent of whether the let- ter appears as a letter or as a number indicating its place in the alphabetical order, a through z. So also with numerically encoded words and sentences. As long as the encoding scheme is known, any numerical symbol, or com- bination of numerical symbols, can be "read" as efficiently as the combinations of corresponding alphabetical symbols. Such interchangeability also exists, though to a lesser extent, in analog representation. For example, the sounds on a phonograph record can be transcribed onto a cassette
14 tape, and two measuring devices can be calibrated so that the position of the indicator on one conveys information about the position of the indicator on the other. One should also note that because information can be represented by two fundamentally different methods, the representation of it in one will have an equivalent in the other. This is particularly true in the computing and telecommunications fields where techniques for analog-to- digital conversion, and vice versa, are necessarily well developed. Finally, it is important to observe that information- conveying mechanisms often have both analog and digital aspects, and, when they do, it can be helpful, and sometimes essential, to appreciate the differences between them. As noted, the automobile speedometer has both a moving needle and a numerical scaleâthe one sufficient to observe accel- eration; the other essential to staying below the posted speed limit. A clock has both circulating hands and a face engraved with sequences of symbols that conveniently refine the information conveyed by variation in the posi- tioning of the hands. A musical instrument is a device for reproducing the discrete tones of the musical scale, which is digital in nature, while the actual playing of the in- strument relies on a form of analog representationâ!.e., sound created by continuously varying the pressure of the surrounding air. For the present discussion, one need not examine in exquisite detail the analog or digital attributes of things that are not central to understanding computing and tele- communications capabilities. In what follows, however, it is important to bear in mind that information as a quantum of knowledge can be represented in many forms, some of which may be analog, others may be digital, and combina- tions of the two and conversions from one to the other also are possible. CONTEMPORARY COMPUTING AND TELECOMMUNICATIONS TECHNOLOGY There abound in the world today two broad classes of com- puting technologyânot surprisingly, one called analog and the other called digital. In the analog device, information is typically represented as continuously varying electrical signals whose direct manipulation allows the performance of desired algebraic or mathematical operations. In the digital computer, on the other hand, information is in the form of sequences of (usually numerical) symbols whose
15 physical representation is an electrical voltage or signal that may have only discrete values or an element of mag- netism that may have only particular positions. The basic computing operations consist of combining such electrical or other signals, according to appropriate rules, such as those of arithmetic, and, because all information is represented and manipulated digitally, encoded nonnumeric information, such as letters, words, and phrases, can be handled as readily as any other. The digital computer, in other words, is in principle capable of accepting and man- ipulating virtually any kind of information, and for that reason, in particular, it has become over the last 15 years the most common electronic computing device. The digital computer's extraordinary versatility in re- gard to the forms of information it will accept also ac- counts for its wide use today in conjunction with other technologies and, most notably for present purposes, in conjunction with telecommunications technology. Any com- puter, be it analog or digital, has four main functions: to accept information of a variety of kinds; to represent and store it internally in appropriate fashion; to manipu- late it as instructed; and to produce a desired information product. In performing those four functions, the computer may combine fragments of information with one another, dis- card other fragments, or create new information, such as by adding or multiplying a series of numbers. In short, the very essence of computing, or of data processing, as it is sometimes called, is to change information in de- sired ways. The essential function of telecommunications, on the other hand, is to transport information from place to place with the understanding that what is delivered to the recip- ient will be the same as what the sender sent. Regardless of the form in which a telecommunications system accepts information from a sender, regardless of how the informa- tion is represented in transit, and regardless of the form in which it is delivered at its destination, the informa- tion content is not supposed to be altered. A telegraphic communication system, for example, may accept handwritten messages whose representation enroute changes repeatedlyâ from keystrokes to electrical signals, to radio signals to and from a satellite, to a printed telegram delivered to the recipient. Similarly, the voice conversation between two subscribers of the telephone system may at one time be represented as analog electrical signals and at another time as digital signalsâeach changing one to the other typically without affecting information content.
16 Notwithstanding these basic functional differences, how- ever, computing and telecommunications, and particularly digital computing and telecommunications, are proving to be extraordinarily useful to one another. Telecommunica- tions can extend the outreach of computing capability by connecting it to remotely located terminals, printers, and displays. By thus allowing the processing power Of one or more computers to be delivered to places of choice in a flexible, economical, and appropriate manner, telecommuni- cations technology encourages the development of new and often highly practical computer applications. For example, a merchant today can use a counter-top terminal to check the validity of credit cards presented as payment for pur- chases; a branch bank can verify that a customer's balance is sufficient to cover a proposed withdrawal; and employees of a government agency, such as the Social Security Admin- istration, can assist clients by querying the agency's central records through terminals located in widely dis- persed field offices. Conversely, digital computing technology can greatly enhance the flexibility and efficiency of telecommunications capability. In the domestic and international telephone net, for example, digital computers are now used to perform such tasks as: â¢ Selecting routing .paths and establishing circuits; â¢ Monitoring the utilization of equipment; â¢ Producing bills; â¢ Accepting information flowing in on one communica- tions channel, storing it temporarily, and then routing it outward on another; â¢ Monitoring the reliability and overall performance of the network; and â¢ Allowing information represented in digital form to be broken up into easily transportable units that are then reassembled at their destination to recreate the original data stream.1 In sum, digital computing technology is playing an increas- ingly significant role in the management, operation, and control of telecommunications systems at the same time that telecommunications systems are facilitating the spread of digital computing capacity. Moreover, this growing syner- gism between the two technologies, sometimes referred to as the marriage or convergence of computing and telecommunica- tions, is being continuously reinforced by technological advances in both areas.
17 Traditionally, computing and telecommunications have been considered to be based on distinct technologies, and in many respects they still are. Each has its own special- ized equipment, for example. A telecommunications system uses microwave relays, satellite transmission facilities, modulating devices of various kinds, and, of course, all of the open-wire and coaxial cable technology. Likewise, a digital computing operation uses magnetic tapes, magnetic disks, various types of printing devices, and some special electronic technology such as magnetic cores. Increasingly common to both, however, and, in effect, encouraging them to move ever closer together, is the exploitation of modern-day solid-state integrated circuit technology. TECHNOLOGICAL PROGRESS The first truly electronic digital computerâENIACâ exploited World War II vacuum-tube technology. The ma- chine, completed in 1946, used some 20,000 vacuum tubes to execute 100,000 or so calculations per second. It had a memory capacity of only 20 words (of 10 decimal digits each), and amply filled a room of more than l,200 square feet. Over the years vacuum-tube technology gave way to miniature vacuum-tube technology, which in turn yielded to the first transistor technology, and eventually to the contemporary densely-packed integrated circuit. As a consequence, a complete microcomputer today, including even memory capacity, can be contained on a silicon chip perhaps 1/4-inch square, and such a tiny electronic mecha- nism can outperform ENIAC hundreds of times over. Furthermore, the density with which basic electronic entities can be placed on such a chip is steadily increas- ing, so rapidly, in fact, that by the mid-1980's a memory chip that can store a million binary digitsâthe equivalent of some 30,000 of the old ENIAC wordsâis expected to be commercially available. Not only has the packing density of such silicon-chip technology been increasing dramatically, but, importantly, the cost of manufacturing microchips has declined impressively. Thus, ever-larger computational capacity is becoming available in ever-smaller physical size at ever-decreasing cost. Computations that were done for tens of dollars three decades ago can now be done for pennies or fractions thereof. To illustrate how dramatic the advance of solid-state electronic circuit technology has been, one recent report noted that if the price-to- performance ratio of a Rolls Royce had evolved in comparable
18 fashion, today, the most luxurious model would cost slightly less than 25 cents.2 (One might also note that it would go a million miles on a gallon of gas!) In many respects, progress in the telecommunications field has been equally remarkable. The capacity of the original copper or iron open-wire line, subsequently en- hanced by various carrier systems, has been supplemented by the broader bandwidth, or carrying capability, of coax- ial cables and microwave relays. Today, enormous numbers of individual voice and other subscriber channels can be "multiplexed" together for satellite transmission, and just entering use is the glass optical fiber, which provides enormous capacity in a very small physical size. Further- more, as the demand for communications links dedicated to data transmission has grown, along with the demand for so- called "dial-up" data communication, the telephone industry, in particular, has been responding with new ways to better use older facilities as well as by providing new capabil- ity with improved technology. Historically, data communicationâi.e., communication between central computers and remotely located terminals or other computersâhas been constrained by the carrying capacity of the installed telephone network. That limita- tion has been due in part to the original design of the telephone system for the bandwidth of voice communication, and in part to the many subsequent improvements that have necessarily been shaped by the statistical properties of subscriber behaviorâe.g., duration of call, frequency of calling, geographical distance between connected parties. For computer-related data transmissions, however, both the optimal bandwidth and the statistical properties of usage are quite different. For example, a particular circuit may sometimes be needed only for milliseconds but at other times continuously for hours. Thus, the designers of telecommunications facilities have had to make special efforts to understand the needs and behavior of data communication subscribers, and, as they have done so, some of the constraints arising from the past emphasis on voice communication have begun to disappear. Some, in fact, have disappeared simply as a result of new digital techniques developed to make more efficient use of voice communication facilities. "Pulse code modulation," for instance, is now a well-established innovation that represents voice conversations in digital form, thereby enlarging the number of them that can be carried on simul- taneously over a given communication link. (As should be clear from earlier discussion, each conversation, while
19 temporarily represented as a digital stream, is restored to its original analog form prior to its delivery to a re- ceiving handset.) Because pulse code modulation is a digi- tal technique, however, its use by the telephone industry inevitably facilitates the industry's ability to transport other types of digitally represented information and par- ticularly the digital data streams generated by computers and computer terminals. Advances in solid-state integrated circuitry are also having an impact on telecommunications, though to date a less direct impact than the one on computing. Modern-day telecommunications systems may have parts that are coaxial or microwave, parts that are satellite-based, and parts that rely on twisted copper wires. This, by itself, limits the opportunity to exploit microelectronics, but, in addi- tion many system components appear to be ones to which the new chip technology may contribute littleâsatellite antennas, for example, and substantial portions of the transmitting and receiving devices. Still, one can confi- dently expect the impact of microelectronics to grow as the telecommunications industry learns how solid-state inte- grated circuitry can be used to improve the efficiency of existing operations, to develop substantially new and sig- nificantly different services, and, most importantly, to satisfy the burgeoning demand for data communications capacity. NEW KINDS OF FACILITIES FOR DATA COMMUNICATIONS Because the demand for data communications capacity is growing by leaps and bounds, virtually all segments of the telecommunications industry are now working on ways to transport information in digital form. Experiments with digital broadcasting, using transmission channels assigned to local television and FM radio stations have been under- way for several years, and just over the horizon is large- scale exploitation of fiber-optic technologyâsometimes called optical cableâwhich uses laser-generated signals instead of electrical ones to transmit digitally represented messages. Total capacity of such an optical cable is many times greater than that of conventional copper-wire cables, and thus can provide an enormously enhanced capability for information distribution, particularly in densely populated areas. A handful of new business ventures have also sprung up, marketing specialized forms of telecommunications systems,
20 such as "packet-switching" networks or, more precisely, "packet data communications." This is a method of digital data transmission that divides data traffic into packets of fixed length, with no limitation on the number of packets that may be used. For example, if a transmission consists of 100,000 digital elements, it might be fractured into 100 packets of 1,000 elements each. Each such packet is then individually addressed so that enroute-switching and control centers know how to handle each one and direct it toward its intended recipient (typically a computer that performs data-processing operations). Because each packet travels through the network as an individual entity, not all will necessarily follow the same path, but all will arrive at their intended destination where they can be re- assembled in proper sequence, thus recreating the original data stream. The packet approach is one that exploits the statistical characteristics of data communication, but, in addition, it is a clear example of digital computing technology being innovatively exploited for a telecommunications purpose. Because all information flowing through such a system is represented digitally, digital computers are readily used to control packet routing, to detect and control errors, and to maintain the integrity of the network if trouble develops in some of its parts. In other words, the digital computers in a packet network act as control mechanisms that see to it that the transmission of information takes place as intended and that the network fulfills its basic obligation as a telecommunications systemânamely, to de- liver information from sender to recipient in a timely man- ner, without errors or other changes in the information content of the original message. Packet switching, it should be noted, is also a good example of how an application of digital computing tech- nology can have significant economic implications. Packet techniques were first implemented as the ARPANET, funded by the Department of Defense in the late 1960's. Con- ceived as a vehicle for sharing computing capabilities among a group of government-funded research centers, the ARPANET now contains some 60 "nodes," each of which is either a computer center providing services to the net or a research operation using services from the net. Such an arrangement makes it possible for the participating re- searchers to share one another's computing installations and also encourages the sharing of computer software, of- ten at appreciable cost saving. At present the packet concept is being commercially
21 exploited. Throughout the world a half dozen such net- works are either in operation or under construction. From the commercial users' point of view, one of packet switching's strong advantages is that it allows the tele- communications carrier offering it to charge users for the amount of traffic they actually generate rather than for the amount of time that toll circuits are committed to them. In the past, because of the technological need to dedicate circuits to users, many paid for time they could not fully use. With the advent of packet switching, such dedication is no longer essential, and thus not only can price be brought into line with actual usage, but previously underutilized circuits can be operated more efficiently. TELECOMMUNICATIONS EFFICIENCY AND OTHER ISSUES Packet services may be supplied by a specialized common carrier or as a specialized service within the framework of the classical telephone system. For the future, an im- portant market for them will be record-keeping operations whose information comes from a variety of geographically dispersed locations. One illustration is the so-called "point-of-sale" electronic funds transfer system in which a purchase would be paid by using a terminal on the merchant's premises to debit the customer's bank account and 'credit the merchant's. In any such transaction, the amount of data generated is small and need only be trans- mitted each time a sales transaction occurs. Hence, while there might be thousands of merchants using such a system, the vast majority of them would not generate enough data traffic to warrant full-time use of a dedicated communica- tion channel, whereas a system based on packet techniques would provide them with as little or as much transmission capacity as they might need. There are, of course, other approaches to implementing a funds transfer system of the sort described. For example, one could aggregate data from many sources for transmission through a full-time communication channel that is collec- tively shared, and, in fact, such an approach is currently being taken by some transaction-oriented systems, such as the reservations systems the commercial airlines operate. One should also note that in order to make a packet system function properly, extra items, such as error control dig- its, address digits, and sequence numbers, must be added to the basic information being transmitted. Thus, there is a
22 form of overhead inherent in packet techniques that, for a user who does have a large quantity of steadily flowing data to transport, may make them less cost-effective than the more conventional dedicated circuits. Indeed, the question of how best to provide telecommuni- cations among geographically distributed terminals and one or more computing centers that service them is clearly an engineering design problem that involves much more than questions of technical feasibility. One would like to think that eventually conventional techniques for managing and using a telecommunications network, plus packet switch- ing and other innovations that are yet to be made, will give telecommunications users all the options necessary to satisfy their requirements in whatever way is most cost- effective for them. To reach that point, though, many obstacles still need to be surmounted. As explained in Chapter III, the influence of government regulation on the future of data communication services is of major importance. In the large, there is ample tech- nology already available for much of the data-oriented telecommunications services that users will demand, and additional technology is constantly emerging from the re- seairch laboratories. From a purely technological point of view, that is, there seems to be no reason why today's telecommunications systems could not be augmented to the point where a user would be able to send and receive vir- tually any kind of information (i.e., print, data, speech, facsimile, full-motion television images) economically and whenever the user needs or wants to do so. As in many other high-technology areas, technical feasibility is only the first requirement. The U.S. telephone and telegraph industry has long been a regulated monopoly, and in recent years regulation, for a variety of reasons, has tended to slow its technological evolution. Furthermore, because the industry must depre- ciate its installed plant over a long period of time, and because the prices that may be charged for a regulated ser- vice are typically set with a view to making the service affordable by small users as well as large ones, there is a reluctance to allow new competitors to enter the tele- communications field if, in so doing, they might siphon off customers the established carriers consider necessary to maintaining their existing base of revenues. In addition to these kinds of knotty regulatory prob- lems, the need for widely agreed upon technical standards could turn out to be a major source of difficulty in some matters. If computers are to "converse" through
23 telecommunications networks easily, interface standardsâ generally called protocolsâwill have to be devised. The task of developing them, however, could turn out to be as contentious as it is complex, since few standards are wholly without competitive ramifications. Finally, there is a formidable group of problems that arise from the applications of computer-based information technology that often stimulate, and occasionally require, changes in the way things have previously been done. Some- times such change is accepted but often it is resisted by people, or institutions, or interests that, for one reason or another, prefer the status quo. Even today one can pro- duce a long list of instructive examples that encompass a wide array of economic, psychological, and, on occasion, political concerns. Consider, for instance, the amount of resistance that some price-conscious consumers showed to the introduction of computerized check-out stands in supermarkets. Quite apart from any consumer resistance, consider also the oppo- sition to electronic funds transfer systems that has devel- oped within segments of the financial industry on the grounds that the systems' remote terminals amount to branch banks and, therefore, threaten to upset the existing pat- tern of intra-industry competition. Consider the public concern that was voiced in the mid-1970's over the use of computer-based information technology to gather, maintain, and disseminate records pertaining to identifiable individ- uals, and note as well the uneven pace of office automa- tion that appears to be due in part to concern that long- established interpersonal or hierarchical organizational relationships will be disrupted.1* Lay these examples along- side the growing controversy over the future role of the U.S. Postal Service as a provider of electronic message servicesâa controversy that involves issues of public ver- sus private initiative and ownership, conflicts among regu- latory agencies with overlapping jurisdictions, and unsolved policy and technical issues relating to the confidential status of first-class mailâand one begins to see clearly why successfully extending the utilization of computer-based information technology requires a solid appreciation of matters far beyond the narrow question of technological feasibility. Then, too, there are in no way farfetched applications of the technology that may contain latent problems of which few to date have even an inkling. For example, small microprocessor-based home computers for which an interface to the telephone system is readily available are growing
24 in popularity. In what ways might they pose new threats to personal privacy? In what ways might their widespread use facilitate subtle incursions by innovative people into public and private information systems that also use the common-carrier telecommunications networks?5 Questions of this sort are multiplying today, and unless they are forthrightly and responsibly examined, there seems to be little doubt that the pace at which computer-based information technology is applied will continue to be slower than its demonstrated usefulness throughout large sections of our society might lead one to expect. The joint appli- cations of computing and telecommunications technology that have been made so far have been extraordinarily rich in servicesâfinancial and reservations services, public data bases, shareable computing power, command and control systems for the military, corporate audit and planning capability, stock and commodity price quotations, individ- ualized education in grammar and arithmetic, traffic-control support for air, water, and rail transportation, weather reporting, scientific data banks, newspaper morgues, spe- cialized services for medical professionals, remote print- ing capability for newspapers and magazinesâand on and on. Nonetheless, the sum of what has been accomplished to date is less than what might have been accomplished if the technology had been recognized as the potentially power- ful technological resource that it is. Further, if we are to begin now to recognize it as such, and also, and most importantly, to recognize it in a way that directs attention to the differences between desirable and undesirable appli- cations of it, the amount of research and analytical effort in support of public and private decision making about the subject needs to be enlarged.
25 NOTES 1. For an explanation of packet-switching techniques and their utility, see below, pp. 20-21. 2. Simon Nora and Alain Minc, L'Informatisation de la Societe, Rapport a M. le President de la Republique, La Documentation Francaise, Paris, 1978 (English translation prepared by Transmantics, Inc. of Wash- ington, D.C. for the National Telecommunications and Information Administration, U.S. Department of Commerce), p. 20. 3. Regulatory delays of several years in each case were cited in note 10, Chapter l, for domestic satellites, mobile telephone, cable television, and hybrid com- puter communications services. To this list may be added regulatory delay in authorizing various transat- lantic telephone cables and "teletext," or alphanumeric information services for home television screen display. Such delays mean that new markets cannot be counted on with certainty, so that incentive lags to develop the innovative services and facilities to supply the markets. Another deterrent to innovation is long de- preciation schedules set by regulation. For example, mechanical No. 5 crossbar switching equipment has been retained in the rate base well beyond the time that it was technologically surpassed by electronic switching. An example of innovation not held back by regulation is optical fiber transmission technology; the decision in this case could be made by the firm within existing regulatory policy. 4. See, for example, U.S. General Accounting Office, Fed- eral Productivity Suffers Because Word Processing is Not Being Well Managed. Washington, D.C.: U.S. Gov- ernment Printing Office, April 6, 1979; see also, The Technical Office. Cambridge, Massachusetts: The Yankee Group, March 1978, p. 138. 5. An alleged case of such an incursion involving a pri- vate school in New York City and 21 computer systems operated by organizations in Canada was the subject recently of an affidavit filed by the Federal Bureau of Investigation in support of a Bureau request for a search warrant. New York Times, May 6, 1980, p. 1.