This series of articles by George Gilder provide some interesting technological and cultural background that helps prepare readers to better understand and place in proper perspective the events relative to the National Data Super Highway, which are unfolding almost daily in the national press. I contacted the author and Forbes and as the preface below indicates obtained permission to post on the Internet. Please note that the preface must be included when cross posting or uploading this article.
The following article, ISSAQUAH MIRACLE, was first published in Forbes ASAP, June 7, 1993. It is a portion of George Gilder's book, Telecosm, which will be published in 1996 by Simon & Schuster, as a sequel to Microcosm, published in 1989 and Life After Television published by Norton in 1992. Subsequent chapters of Telecosm will be serialized in Forbes ASAP.
According to Web-Counter, this article has been accessed times since Nov 3, 1995.
In the spring of 1989 when Michael Bookey first visited the Middle School in Issaquah, Wash., to help the school system with its computers, he was reminded of his early ventures into Communist China. After 20 years of working with computer networks, to enter Issaquah seemed to me like encountering an exotic tribe of primitives untouched by the modern world. The only sign of modern technology was a forlorn computer room full of Radio Shack TRS-80 machines, most of which had broken down. Then he learned that as a remedy for this problem, the district had recently voted a levy of $2.7 million for outlays on high technology. Lacking any better ideas, the school system had decided to distribute the money equally among the teachers, to spend as they wanted. What they wanted turned out to be VCRs, incompatible CD- ROM drives and a random selection of computers, printers and other gear to be scattered through the schools under the influence of a flock of computer salespeople attracted to the site by the pool of mandated money. To Bookey, this remedy seemed worse than the disease. It meant that the bulk of the money would be wasted, further estranging both taxpayers and students from the most powerful technologies of their era. Bookey wanted school officials to know that the most powerful technology is not computers, but computers joined in networks. Explaining the magic of networks, Bookey asks you to imagine a car plumped down in the jungle. Checking it out, you might find it a very useful piece of equipment indeed. A multipurpose wonder, it would supply lights, bedding, radio communications, tape player, heat, air conditioning, a shield against arrows and bullets, and a loud horn to frighten away fierce animals. In awe of the features of this machine, you might never realize that the real magic of a car comes in conjunction with asphalt. For the first 10 years of the personal computer era, according to Bookey, we have used our computers like cars in the jungle. We have plumbed their powers for processing words and numbers. All too often, home computers have ended up in the closet unused. We have often failed to recognize that most of the magic of computing stems from the exponential benefits of interconnection. In the microcosm, the interconnections come on individual chips, as ever smaller transistors crammed ever closer together work faster, cooler and cheaper, enhancing both the capability and the speed of the processor. The microcosm strewed some 100 million personal computers around the world and endowed individuals at workstations with the creative power of factory owners of the Industrial Age. Just as the microcosm generates exponential gains from increasing connections on chips, the telecosm generates exponential gains by increasing connections between chips, powerful microcomputers in themselves. These links between increasingly potent microchips will soon dominate the world of communications. The networking industry therefore faces a drastic transition from a people-to-people regime to computer-to-computer. This change is so radical that it resembles a mutation that creates a new species. People communicate in domains of time and space entirely alien to the world of computers. To a person, a one- second delay on a voice line seems hardly noticeable; to a computer, one second may mean a billion computations that would take hundreds of human lifetimes to accomplish by hand. Most important, people can transmit or receive only a small stream of information at a time. They want relatively narrow bandwidth connections for a relatively long period, a 64-kilobit- per-second voice link, for example, for a 10-minute phone call. Computers, on the other hand, can handle hundreds of millions or even billions of bits a second. They often need many millions of bits of bandwidth for a short time fractions of seconds. As industry shifts from a human scale of time and space to a computer scale, the systems and structures in existing telephone and broadcast networks become almost irrelevant. Essentially, all other forms of networks: voice, text, video and sound, are rapidly giving way to various new forms of multimedia computer networks. Driving this overwhelming force of change is the alchemy of interconnections, working in the telecosm with the same logic and feedback loops as connections in the microcosm. hile dumb terminals such as phones and TVs use up bandwidth without giving anything back, computers are contributors to bandwidth, not consumers of it. In general, the more computers, the more bandwidth. Not only is the network a resource for each new computer attached to it, but each new computer is also a resource for the network. Each new computer expands the potential switching and processing capacity of the system by a large multiple of the increasing demands it makes on other switches and processors. As ever more powerful computers are linked ever more closely, whether in digital cellular microcells or in webs of fiber and coaxial cable, usable bandwidth expands explosively. Governing the expansion of networks, the law of the telecosm is just as potent as the law of the microcosm. Indeed, in enhancing the productivity of organizations, the telecosm consummates the microcosmic miracle.
The creator in the early 1970s of what may have been the world's first fully functioning system of corporate electronic mail, Bookey was quick to foresee this radical shift from person- to-person to computer-to-computer communications. Pursuing his vision of networks, Bookey in 1982 spurned a possible job at Microsoft on the grounds that the company was outfitting cars for the jungle, a decision that probably cost him several million dollars. Instead, he joined Seafirst Bank in Seattle, where he made history (in the form of a reference in John Sculley's autobiography, Odyssey) by pushing the purchase of a thousand Macintosh computers for bank networks at a crucial time for Apple. In 1986 Bookey left the bank to join Doelz Co., a startup in Irvine, Calif., that built advanced computer network equipment that he had used at Seafirst. For Doelz, Bookey designed software and spearheaded marketing. A so-called cell-based network, the Doelz system broke up a stream of data into short, equal-sized packets, each with its own address, to be sent through the nodes of the net in nanoseconds, like letters accelerated a trillionfold through the branches of the post office. Bookey was not necessarily wrong in choosing this technology over Microsoft's. In the form of asynchronous transfer mode (ATM) systems, this essential approach, based on short, uniform packets that can be switched at gigabit speeds in hardware, is now the rage of planners in the computer networking industry. ATM is seen as the crucial enabler for digital networks combining voice, data and video in so-called multimedia applications. Bill Gates now calls multimedia the future of his industry. Although many observers still see ATM as a futuristic technology, Bookey believes its future is nearly now. From the humblest personal digital phone to the most advanced supercomputer, computer-to-computer links will dominate the entire universe of telecommunications, and ATM will dominate network switching. Doelz, however, was ahead of its time and failed to survive a tangled legal imbroglio with AT&T in 1988. So Bookey took a big profit on his California residence and returned with his wife Robin and daughter Erin to Seattle, where he had grown up and set records in the mile on the track at the University of Washington. He bought his dream house on the top of Cougar Mountain in Issaquah, with a view of the very Twin Peaks made famous in the television series and put out his shingle as a network consultant under the name Digital Network Architects (DNA). Almost as an afterthought, the Bookeys sent Erin to Issaquah Middle School. Having designed networks around the world, Bookey had often seen their powerful impact on business organizations, such as banks. Bookey believed that networks could have a similar revitalizing impact on schools. Like banks, schools are essentially information systems that have brought their Industrial Age hierarchy into the Information Age. Creating networks in schools, however, posed many special problems. Most school systems, like Issaquah, were largely unaccustomed to managing technology. The system would need to create a large MIS (management information services) organization just to keep the network functioning. Then, as the teachers at Issaquah hastened to point out to Bookey, there was the problem of students. Impulsive, mischievous and messy, they in no way resembled the disciplined employees of a corporation. Speaking from grim experience, some of the teachers told Bookey that his network plans would succeed only if the computers were reserved exclusively for teachers and if students were barred entirely. Bookey, however, thought there had to be a way to bring the magic of networks to America's increasingly troubled school systems. The secret would be to recognize that, just as computers are not consumers of but contributors to bandwidth, students should be seen not as a problem, but as a precious resource in launching the networks that inform the Information Age.
Ever since Adam Smith first maintained that the division of labor, the spread of specialization, is the catalyst of the wealth of nations, economists have seen the breakdown of functions into subfunctions and specialties as the driver of efficiency and growth. The key force expanding specialization in the contemporary capitalist economy is networks. Indeed, networks, by their nature and purpose, refine the division of labor. In the financial industry, for example, networks allowed the proliferation of specialized institutions. In the ever-shifting kaleidoscopes of American finance, some institutions went local, some global. Some managed car loans, credit cards or other consumer services; some handled mortgages, mutual funds or real estate trusts; still others stressed computer leases, junk bonds, venture capital or large corporate accounts. The pell-mell fragmentation of American finance during the 1980s into an ever more refined division of labor enabled the U.S. to lead the world in levels of capital efficiency, with more economic growth per dollar of savings than any other country. Each financial business did not have to repeat all the work of all the rest, and each became more efficient at a particular task. Bookey believes that networks can have a similar effect on that other great information-processing industry: education. Why should every school have an all-purpose library and a French teacher and a calculus scholar and a health center and an administrative office? Why should every school have an entire complement of buildings? With all the schools on networks, individual schools could specialize in particular subjects, functions and resources, as financial companies do. Education would not have to happen exclusively, or even mostly, in schools. The explosive spread of networks is now the prime mover of the U.S. economy, allowing all industries to break down into patterns of specialization unbound by place and time. And now the government wants to get into the act.
Zoom through tax-hike tollgates and glide out onto data superhighways; this is the new mantra of American industrial policy. Add the further fillip of investment for educational infrastructure and you can sweep up the ramp toward the federal treasury and drive out with a bonanza. In this new era of the big bands, there are now some 10 bills before Congress to foster vast new networks with large bandwidth, or communications capacity. Some $2 billion has already been authorized and $765 million appropriated this year for various programs related to a National Research and Educational Network (NREN). Candidate Bill Clinton presented the concept of NREN as Ra national information network to link every home, business, lab, classroom and library by the year 2015. President Bill Clinton, vice-president Albert Gore and a raft of advisors all celebrate the highway as the metaphor for the future information economy. Gore points out that his father was a leader in building the Interstate Highway System in the early 1950s; Albert Jr., wants to play a key role in building the information highways of the 1990s. Indeed, data superhighways would seem to be the fulfillment of the fibersphere; the way to create the vast new infrastructure of fiber-optic lines that will bring the full promise of digital video and multimedia communications to all citizens. Why, then, is Mike Bookey so worried? He would seem to be the perfect NREN champion. Bookey has pursued networks through most of his career and now is focusing on networks for education. In explaining the importance of computer connections, he has even long used Gore's favored highway metaphor. Bookey thinks that the federal superhighwaymen do not grasp the nature of networks and how they grow. In systems work we have a rule: You design top down, but you build bottom up. Bookey sees the creation of networks as an organic process, driven by public demand, shaped by human needs and rooted in a moral universe of growth through sharing. It is the experience of building the network that creates the expertise to maintain and use it. In all these processes, big government is nearly irrelevant.
For the past 10 years, Washington, D.C. experts have been wringing their hands over the supposedly unbearable costs of building broadband networks and the urgent need for large federal funding. Analysts have been ruminating over the question of who would spearhead the creation of broadband nets; the phone companies, the cable television companies or the government. Before any of these forces could act, however, it became clear that the answer would be none of the above. The hardest part of the job was accomplished, with astonishing speed, by computer and networking companies. The rest of the work is well under way, as cable and phone companies adopt the computer technologies. As recently as 1989, only seven percent of America's personal computers were connected to local area networks. By 1991 45 percent were connected, and by 1993, close to two-thirds were linked to LANs. Growing even faster than LANs is the internetworking business: the interconnection of existing local area nets in wide area networks. Building internetworking gear or accessories, such companies as Cisco Systems, Cabletron, Wellfleet, 3Com and SynOptics are among the highest flyers in the technology stock market boom. Cisco, for example, is growing some 50 percent a year and commands a market value of almost $6 billion, comparable to that of Digital Equipment Corp. Cabletron has hiked its revenues some 16-fold in the last five years. Most of these connections run at some 10 megabits per second, enough for high-resolution digital video, but inadequate for the more exotic traffic in images predicted for use later in the decade. Increasingly, however, the connections are fiber- optic lines or are broadband coax, which is nearly as good as fiber for short-distance transport. The potential of fiber is almost unlimited (see "Into the Fibersphere," Forbes ASAP, December 7, 1992). Although moving more slowly than the computer firms, telephone and cable companies are rushing to lay fiber ever deeper into the nation's neighborhoods. Spending some $2 billion (as much as NREN), Telecommunications Inc. (TCI) vows, according to CEO John Malone, to have 90 percent of its subscriber households served by fiber to the curb by 1995. Bringing fiber into the local loop at a slower pace, the telephone companies, led by Bell Atlantic, also are forging ahead with ingenious new ways to make their twisted-pair copper connections carry as much as six megabits per second of digital information. Wireless technology is also moving into the local loop for video delivery (see "The New Rule of Wireless," Forbes ASAP, March 29, 1993). The U.S. networking industry is not in need of fixing. The U.S. currently commands some three-fourths of all the world's LANs and perhaps 85 percent of its internetworks. Although Gore and others justify their industrial policies by referring to the imperious plans of Japan, the U.S. currently commands about three times the computer power per capita as Japan, some 10 times as many computers attached to networks, and an installed base of broadband fiber and cable nearly 10 times as large. The remarkable thing is that the U.S. government is so eager to fix a fabulously flourishing system that is the envy of the world. The electronic and photonic networking industries actually resemble highways in only the most superficial way. The highway construction trade has not advanced substantially in 50 years. By contrast, the networking trade is the fastest-moving part of the ever-accelerating computer industry and doubles its cost- effectiveness every year. Although interconnecting government laboratories, contractors and supercomputer centers with fiber is desirable, a massive government network is not. Issaquah offers better guidance for the future.... But first it will be necessary to deal with the abiding menace of the student problem.
"What do you think you are doing? Answer me," the voice insisted with the I've-got-you-squirming-now confidence of a teacher who has caught a pupil red-handed. "Just lookin' around," grumbled Lee Dumas, the red-headed 13- year-old, trying to sound natural. Glimpsing a telltale red screen of network management among the array of blue displays used in the keyboarding class, the teacher had walked up silently behind Dumas as he broke into the student lists, software programs and grades, and was on the verge of entering the administrative server. Dumas was a bad kid. No one at Maywood Middle School (one of the 16 campuses in Issaquah) doubted that. His teachers called him "obnoxious" or even "brain-dead." He set what he believes was an all-time record at Maywood by being detained after class some 60 times for insubordination. Using the approved psychobabble, he says, "I had problems with authority. I couldn't accept teachers ordering me around." After being caught breaking into the computer system, Dumas was dragged up to the principal's office. Neither the teacher nor the principal could figure out the nature of the crime or judge its seriousness. For help, they summoned Don Robertson, the administrator assigned to Issaquah's Technology Information Project (TIP). He considered the situation gravely and recommended severe punishment. Toward the end of the meeting, however, he turned to Dumas and said, "With your talent, you should become the sheriff rather than the outlaw. Why don't you come down and join TIP?" Since no one had previously detected any talent in Dumas, this comment made a sharp impression. About a week later, he showed up sheepishly at Robertson's door. To school administrators, kids like Dumas might be a problem, but to Bookey, Issaquah's 9,000 students seemed a wonderfully cheap resource. By training the students to build and maintain the networks, he could make the $2.7 million the foundation of an enduring educational resource. In the end, the Issaquah network was almost entirely built by students between the ages of 12 and 17. Using students to solve the problems of network maintenance and support and thus reduce the real costs by some 80 percent was Mike Bookey's solution to the perplexing problem of computers in schools. The first step in the Issaquah networking venture, in the spring of 1990, cost no money and arose from pure necessity. Just as in businesses across the country, the initial motive for networking was the arrival of laser printers from Hewlett Packard. Bookey began by giving his 10-person TIP team a pile of manuals and having them install a basic network connecting two PCs, an Apple II and a Macintosh to a laser printer. This step enhanced the value of all the computers at a small fraction of the cost of buying new dot-matrix printers for each. Four of the ten students managed to cobble together the network in about a month. They learned the intricacies of pulling twisted-pair wiring for 10baseT Ethernet computer connections running at the standard rate of 10 million bits (megabits) per second. The next step was to add a hard disk containing school files and software programs. Using both Apples and IBM PCs, the Issaquah network from the beginning, had to handle a variety of communications protocols. If the network was to connect to anything outside itself to the school's administration building or the school system's libraries, for example, Issaquah would have to install equipment that could sort out messages from different computers. This meant Issaquah joined the market for multiprotocol routers. A router is a device that sits on a computer network and reads the addresses on all the message packets that pass by. If the address is on another network with a different protocol, the router creates a new envelope for the packet and sends it to the other network. Nonetheless, with all their routers and Ethernet wiring, the Issaquah networks slowed to a crawl as soon as they had to connect outside a building. There, they had to depend on what is known as the Public Switched Telephone Network, where everything turns to analog and drowses down to some 2,400 bits per second. Bookey demonstrated that the school could save money on its voice communications by buying a digital T-1 line that multiplexes 24 phone circuits onto a 1.544-megabit-per-second system. Since 12 of the 24 circuits would be enough to satisfy the school's internal voice needs, the rest of the T-1 line, some 760 kilobits per second, could be devoted to the data communications needs created by the school's new Ethernets. Thus, while getting a cheaper solution for its voice traffic, the school increased its data bandwidth by some sevenfold for free. Once these connections were in place, the students acquired a Microsoft Mail program to incorporate E-mail in the system. Soon, this became the heart of the network, with both students and teachers using it constantly to handle papers, consult teachers in other schools in the system, make reports to the state and interact with parents and students. E-mail became so central to the functioning of Issaquah that when the computers were down teachers would talk of canceling classes. To E-mail were added connections to Internet, the global research and education network launched some 33 years ago as DARPA Net (the Pentagon's Defense Advanced Research Projects Agency). Since Internet was civilianized in 1983, adopting the TCP/IP networking standard, it has been expanding its traffic at a pace of some 15 percent per month. Between 1981 and 1992 the number of computers connected to Internet rose from 281 to 1.1 million. Through Internet, the students could search through a variety of databases for material for a paper or connect to Japan for help in learning Japanese. Along with several other Issaquah students, Aaron Woodman, Jr., a burly boy with his long blonde hair in a ponytail, became so adept at using Internet that he now gives speeches to national conferences on the subject. The speechmaking needs that grew out of the Issaquah project have imparted valuable lessons in English communications for the students. All these developments did not occur without administrative resistance. But the administration eventually became a prime beneficiary. Soon, the computer networks in the Issaquah system were connected by a T-1 line to the Washington Schools Information Processing Cooperative (WSIPC) 20 miles north in Redmond, where attendance and other student records were kept for the entire state. To make these WSIPC services more readily available to schools across the state, Bookey proposed the creation of a statewide educational network running on T-3 lines (45 megabits per second), now known as WEDNET. This provides links all over Washington, from Shaw Island and Stehekin to Seattle and Issaquah, with a rogue line down to Portland, Oreg. As for Lee Dumas, according to his mother, his situation has changed completely, "both in his attitude toward school and in the school's attitude toward him." After joining TIP, Dumas became one of its most active and enthusiastic members. Last summer, he got a job at the Computer Store in Seattle teaching the Macintosh HyperCard program to a student body consisting, yes, of public school teachers. According to Dumas, they had no problem accepting his authority as a fledgling computer guru. No longer one of the outlaws, Dumas became an official beta tester for the new Microsoft DOS 6.0 and Windows NT operating systems, specializing in their security procedures. Following the path of another student who found the "Issaquah bug" in Microsoft's LAN Manager program, Dumas believes he found three or four bugs in NT. Having just finished his sophomore year, Dumas has gone to work this summer at Microsoft for the company's network development chief, Brian Valentine, who regards this once brain- dead punk as a valued employee with high promise for the future. This student who floundered in the usual educational system flourished when his individual specialization was discovered. The Issaquah economy released his energies, just as the national economy releases its own energies through the specialization and division of labor in computer networks. Since there are millions of Lee Dumases in the schools of America, many of them being given up for lost by analysts such as Labor Secretary Robert Reich, because they are not adept at the usual curriculum for "symbolic analysts," Dumas' redemption by technology bears crucial lessons. The lessons are Bookey's: Students are a resource, not a rabble; specialized practical experience is more edifying than most textbook learning; networks are the critical technology both for economic growth and for educational renewal. To these insights should be added Lewis Perelman's view, in his book "School's Out" (1992, Morrow), that teachers should increasingly abandon their role as a "sage on the stage" in favor of service as a "guide on the side," steering their students through a global cornucopia of educational resources.
It may seem peculiar that Bookey, a network guru for large corporations like U.S. West, should focus his attentions on such problems as interconnecting school children in Issaquah with libraries in Bellevue, parents on Squaw Mountain, teachers across town and administrators at the Washington State Information Processing Cooperative. Yet Bookey believes that the educational application may well drive the creation of a true national infrastructure of digital networks. The networking problems of schools closely resemble the networking problems of a nation full of diverse systems. To achieve their full promise, school networks must link computers of many varieties owned by parents, students and teachers, to administrative servers owned by state and local governments, to printers, libraries and databases. School networks must connect LANs to IBM SNA (Systems Network Architecture) links, to a variety of telephone technologies, from T-1 lines of 1.5 megabits per second to T-3 lines at 45 megabits per second and, soon, to ATM switches and other potential gigabit systems. In all its dimensions, including an acute financial constraint, this challenge is altogether as difficult as interconnecting supercomputers over fiber in an NREN. Bookey relished this challenge at Issaquah. Advocates of NREN might disparage Issaquah as a relatively low-grade network. After all, it currently has no fiber outside of the fiber links in the telephone network that it uses. Without fiber, the network will not be able to accommodate collaborative learning in multimedia forms across the country. Bookey demurs. Buying a fiber-optic network before personal computer technology can manage broadband flows of data is premature. In five years, fiber-optic links will probably cost about one-fifth of what they cost today. When the network is needed, Issaquah will be able to purchase it and, more important, also use it. Moreover, TCI recently offered to install fiber throughout the Issaquah school system for nothing as part of its general program of fiber to the curb. The fact is that big-band technology will come to Issaquah in due course, with or without NREN money. Critics, of course, will carp that Issaquah is a special case "a relatively rich community" that could afford to levy $2.7 million for technology. Yet the Issaquah example is galvanizing schools across the state of Washington and even in California and Arkansas, where Bookey and his colleague Mason Conner have been consulting with education officials. Emulating Issaquah, other districts in Washington have since raised some $140 million for network ventures.
The lesson of Issaquah is that data highways and superhighways, driven by the convergence of microcosm and telecosm, are indeed emerging in America, and at an astonishing pace. They already are revitalizing the economy and society, and are helping to reform the system of education. The only federal initiatives that will significantly assist the process are lower taxes, accommodation of Internet growth and use, and further deregulation of telecommunications. Communication must begin locally, with access to the community. From these local roots can emerge the great branching systems that can interconnect an information economy. By starting from the top, the government risks paving over the pullulating fabric of networking enterprise with a glass ceiling of expensive and misplaced fiber. In 1993 an estimated 37 million personal computers will be sold worldwide. The same forces that impelled the networks of Issaquah will drive the owners of these new PCs to interconnect them to other networks and will induce the owners of the networks to link them together. As the centrifugal force of the microcosm, multiplying and distributing intelligence through the world, converges with the integrating power of the telecosm, the exponential miracles of specialization and growth will gain new momentum. How far can this spiral reach? Internet will soon approach some interesting limits. According to International Data Group, the number of users has risen from 9,800 in 1986, all in the United States, to 4.7 million around the world today. At this pace, Internet will embrace the entire world population by the year 2001. That's one limit. As the system's trunking backbone rises to 45 megabits per second on T-3 lines and above, the sky is the limit for the amount of message traffic. In the first month after the enlargement to T-3 lines in October 1992, usage rose from 3.5 trillion bytes to 4 trillion bytes. All these networks are dominated by text and still pictures. But the miracles of Internet and Issaquah are about to be joined with a new miracle of growth in digital video connections in the local loop.
How soon can this happen? Advocates of NREN speak of this technology being consummated in 2015. But to most politicians and businessmen, a projected date more than five years ahead is essentially a synonym for never-never land; a way of saying, "Forget about it. I'll be retired." The fact is that a widespread system of two-way broadband networks reaching most American homes, schools and offices is less than five years away. All U.S. business planners must come to terms with this transforming reality. Announcements this spring from leading cable, telephone and computer companies; from TCI and U.S. West to IBM and Silicon Graphics; bring the shape of this network into clear focus. Exemplary among plans announced by a variety of firms is Time-Warner's projected system in Orlando. As described by Jim Chiddix, the company's college-dropout technical guru, the Time- Warner showcase venture will be a giant client/server computer network, suggestive of the arrangements now ubiquitous in corporate computing. The wires will be a combination of fiber to the curb and coax to the home. Much of the system's hardware and software will be supplied by computer companies (allegedly including IBM and Silicon Graphics). The "client" computers will be digitized TVs or teleputers linked to powerful database computers that use a parallel-processing architecture to access hierarchical memory systems, from DRAM caches to optical disk archives. These memories will contain terabytes (trillions of bytes) of digital video movies, games, educational software and other programming. Perhaps the most dramatic breakthrough, though, will come in the switches. While much of the computer and telephone world continues to dither about the future of ATM (many consigning it to the pits of 2015), Time-Warner is committed to installing ATM switches, built by AT&T, beginning next year in Orlando. The ATM system will allow Time-Warner to offer telephone, teleputer and multimedia services together, as soon as the regulators allow it. Chiddix predicts that ATM will soon gravitate to local area networks and ultimately become ubiquitous. But the most portentous announcements of all have come from the telephone companies, who have the most to lose from this cable-oriented network design. Both U.S. West and Pacific Bell have disclosed that they are adopting a combination architecture of fiber and coaxial cable closely resembling the Time-Warner and TCI projects. This unexpected action by two leading Baby Bells, of turning their backs on their millions of miles of twisted-pair copper wires shows both the boldness of the new telephone company leadership and the imperious power of this digital technology. From all sides the telecommunications and computer industries are converging on one essential configuration of advanced parallel-processing hardware, client/server database software and ATM switching. As microcosm and telecosm converge in the living room, with interactive digital video and supercomputer image processing, the leading edge of the digital revolution moves from millions of offices toward billions of homes. Just as Michael Milken, then of Drexel Burnham Lambert, and the late William McGowan of MCI in 1983 rescued long-distance fiber optics from the never-never lands of the year 2015 to which AT&T had consigned it, John Malone of TCI, Gerald M. Levin of Time-Warner and Richard D. McCormick of U.S. West in 1993 have burst open the floodgates for fiber and ATM in the local loop. Again, the force behind this revolutionary development was fierce business and technical rivalry in the marketplace. In the real world the ruling principle of network development is not imposed standardization by government but spontaneous order. It springs from the interplay of human creativity and entrepreneurship with the inexorable laws of physics and technology. These dynamics of interconnection in the Information Age will continue well into the next century. The microcosm will yield chips containing billions of transistors, equivalent to scores of supercomputers on single slivers of silicon. The telecosm will yield bandwidth exploding into the terahertz of all- optical networks and the gigahertz of millimeter waves in the air. Provided that rulers and regulators do not stifle this spiral of opportunity, the human spirit "emancipated and thus allowed to reach its rarest talents and aspirations" will continue to amaze the world with heroic surprises. The Issaquah miracle of Mike Bookey and Lee Dumas and all the others, and the continuing miracle of American networks, which was entirely unexpected by the world, will repeat themselves again and again in new forms of entrepreneurship and technology.