11. Heroes and Villains of the Globalization Era: The Internet Revolution: The World Goes Online

My Name is Paul Baran: Building a Network Without a Center

I was born in Grodno, Poland, in 1926, and came to the United States with my family when I was two years old. I grew up in Philadelphia during a period when radio, electronics, and computing were changing rapidly. I was interested less in what a machine looked like than in why it worked—and, perhaps more importantly, why it sometimes failed. That question would follow me throughout my career.

Learning to Think Like an Engineer

I studied electrical engineering at Drexel Institute of Technology and graduated in 1949. Soon afterward, I worked on UNIVAC I, one of the earliest commercial electronic computers in America. Computers were enormous machines then, expensive and rare, and there was no expectation that ordinary people would someday have several of them in their homes. I later worked at Hughes Aircraft while completing my master's degree at UCLA. By the time I joined the RAND Corporation in 1959, I had learned something fundamental about engineering: a system should not merely function when everything goes correctly. A good system must continue functioning when something goes wrong.

A Network Without a Center

At RAND, I began thinking about military communications during the Cold War. The American telephone system depended heavily upon centralized switching facilities. From an engineering standpoint, that created an obvious weakness. Destroy an important central point, and large portions of the communications system could fail. I asked whether we could design something fundamentally different—a distributed network with many interconnected nodes and no single point whose destruction would bring everything down. The inspiration was not a desire to create websites, social media, or electronic shopping. Those things did not exist. The problem was reliability.

Breaking a Message Apart

My solution was to stop treating a message as one object that had to travel along one predetermined path. Instead, a digital message could be divided into small standardized blocks. Each block could move independently through whatever route was available, passing from node to node until reaching its destination, where the complete message could be reconstructed. If one route disappeared, another could be used. Every node would have similar responsibilities rather than depending upon a central master station. British scientist Donald Davies independently developed related ideas and later introduced the word “packet,” which became the name history remembered.

An Idea Before Its Time

I eventually produced a detailed series of RAND reports describing an all-digital distributed network capable of carrying voice and data across the nation. But designing a system on paper and constructing it are two different accomplishments. RAND did not build what became the Internet in 1960, and I never claimed that I alone invented it. My work supplied one of its important foundations. Other researchers—including Leonard Kleinrock, Donald Davies, Larry Roberts, Bob Kahn, Vint Cerf, Jon Postel, and many others—would solve different pieces of the problem. ARPANET appeared later, and eventually separate networks learned how to communicate with one another.

Watching the Networked World Arrive

I left RAND in 1968 and later helped establish the Institute for the Future before creating and working with companies involved in communications, packet technologies, satellites, wireless networking, and other emerging systems. Decades after my original research, computers were suddenly communicating across enormous interconnected networks. By the 1990s, millions of ordinary people were sending email and exploring the World Wide Web. The problem I had studied had changed completely in scale, but the underlying engineering principle remained familiar: information could be divided, routed through a distributed system, and reassembled at its destination.

I Was One Part of a Much Larger Story

I received considerable recognition later in life, including the National Medal of Technology and Innovation, but I remained a researcher at heart. The Internet was not the invention of one man, one corporation, one university, or one government agency. It was the result of many people solving different problems and then discovering that their solutions could work together. I died in 2011, having lived long enough to see distributed networking move from pages of calculations and diagrams into the daily lives of billions of people. If I must apologize for anything, perhaps it is that researchers sometimes make revolutionary ideas sound terribly ordinary. To me, the question was simply: if one path fails, why shouldn't the information find another?

Paul Baran, RAND, and the Idea of a Decentralized Network - Told by Paul Baran

In 1960, I was working at the RAND Corporation in California, looking at a problem that had nothing to do with websites, social media, or online shopping. None of those things existed. My concern was much more severe: how could the United States maintain communications if a major attack destroyed important parts of the telephone network? At the time, communications depended heavily on centralized switching points. From an engineering standpoint, that was dangerous. If too much depended upon a few central locations, destroying those locations could cripple the entire system.

What If There Were No Center?

I began asking a different kind of question. What if a communications network did not have one central point at all? Imagine a web made of many interconnected switching stations. Instead of every message being forced through a central headquarters, information could move through whichever path remained available. If one station failed, the message would simply find another route. If several failed, the network could continue operating so long as enough connections remained. I called this a distributed communications network, and its strength came from something that seemed almost contradictory: losing pieces of the system did not necessarily destroy the system itself.

Breaking Messages Into Pieces

The next problem was determining how information should travel through such a network. Sending an entire conversation along one fixed circuit, as traditional telephone systems did, was inefficient for the kind of system I had in mind. Instead, I proposed converting information into digital form and dividing it into small standardized message blocks. Each block could travel independently from node to node, taking whatever route the network determined was available. At the destination, the blocks could be placed back into the correct order. Years later, the word “packet,” independently introduced by British researcher Donald Davies, became the familiar term for this approach.

The Network Thinks for Itself

I wanted each switching node to make its own routing decisions. There would be no central computer constantly ordering every other machine where to send information. Each node would examine the condition of the network around it and determine a reasonable direction for the message. That meant the system could adapt when routes became congested or disappeared. Today, this may sound obvious because digital networks routinely redirect information around failures. In the early 1960s, however, it represented a major departure from traditional communications engineering.

Putting the Idea on Paper

RAND gave me the opportunity to examine the concept in great detail. Between 1960 and 1964, I developed simulations, diagrams, mathematical studies, and engineering proposals for the distributed system. In 1964, RAND published my work in a series titled On Distributed Communications. The reports examined routing, reliability, digital switching, network structures, security, and methods for transmitting information. This was not simply a sketch of an interesting idea. I wanted engineers to be able to examine whether such a system could actually be constructed.

Why the Telephone Companies Were Skeptical

Not everyone immediately welcomed the concept. The established telephone system had been built around highly reliable centralized switching and dedicated circuits. My proposal deliberately accepted something that traditional engineers often tried to eliminate: individual components could fail. The system would survive because it had redundancy. To me, that was the point. Building every component so perfectly that nothing could ever fail was unrealistic. Building a network capable of surviving failure was a much more interesting engineering problem.

What RAND Did—and Did Not—Build

This distinction matters. RAND did not construct the Internet in 1960, nor did I build a nationwide operational distributed computer network. What we developed was the engineering concept, analysis, and proposed architecture for such a system. Other researchers were approaching related problems independently. Leonard Kleinrock was developing mathematical theories for data networks at MIT. Donald Davies later developed packet-switching ideas in Britain. ARPA researchers eventually built ARPANET. Decades of work by many engineers would be necessary before the modern Internet emerged.

A Foundation for What Came Next

The importance of my work was not that everything that followed copied my design exactly. Technology rarely develops that neatly. The larger contribution was demonstrating that digital communications did not have to depend upon a rigid central structure or a dedicated path between two people. Information could be divided, routed dynamically through a distributed network, and reconstructed at the other end. Once that principle was understood, engineers had an entirely different way to think about communications. What began as a Cold War problem about surviving damaged telephone systems helped establish ideas that later became fundamental to computer networking—and eventually to a world in which billions of machines could communicate without depending upon one central switch.

My Name is J.C.R. Licklider: Imagining a World of Connected Computers

I was born in St. Louis, Missouri, in 1915, and long before I became involved with computers, I studied psychology, mathematics, and the way human beings process information. I was fascinated by how people think, communicate, and solve problems. Computers interested me not simply because they could calculate quickly, but because I began to wonder whether humans and machines might someday work together in ways neither could accomplish alone.

From Psychology to Computing

I studied at Washington University in St. Louis and eventually earned a doctorate in psychology. During World War II and the years that followed, I worked in fields involving acoustics, human perception, and communication. At MIT, I became involved with research surrounding computers and interactive systems. At the time, computers were usually treated as enormous calculating machines. Users often prepared instructions, handed them over, and waited for results. I believed that arrangement was far too limited. A computer should become something a person could interact with while thinking through a problem.

Man-Computer Symbiosis

In 1960, I published an article called “Man-Computer Symbiosis.” I argued that computers and people should form a cooperative partnership. Human beings are particularly good at setting goals, recognizing patterns, asking questions, and making judgments. Computers are particularly good at calculations, repetitive operations, searching through information, and performing tasks quickly. I imagined systems in which the two could work together almost continuously rather than communicating through stacks of punched cards and long delays.

A Galactic Network

By 1962, my thinking had expanded beyond individual computers. I began writing about what I called a “Galactic Network,” a collection of interconnected computers through which people could obtain programs and information from almost anywhere. The phrase may have sounded ambitious, but the principle was straightforward: if useful information existed on one computer, why should a researcher sitting at another computer be unable to reach it? I imagined a network in which geography mattered far less than it had before.

Taking the Idea to ARPA

In October 1962, I became the first head of the computer research program at the Advanced Research Projects Agency, or ARPA. My task was not to personally construct a national network. It was to find talented researchers, support their work, and encourage them to think beyond the limitations of existing machines. I helped establish research relationships involving institutions such as MIT, UCLA, and Bolt Beranek and Newman, and I promoted work on time-sharing, interactive computing, graphics, and networking. Many of the researchers who followed—including Ivan Sutherland, Bob Taylor, and Larry Roberts—carried these ideas much further.

Making Computers Available to People

One of the ideas I strongly supported was time-sharing. Instead of one person monopolizing an expensive computer while everyone else waited, many people could interact with the same machine through terminals. Critics argued that sharing computer resources this way was inefficient. I believed the important question was not simply how efficiently a machine used its processor, but how effectively people could use the machine. Projects such as MIT's Project MAC helped demonstrate that interactive computing could become practical.

Watching the Network Take Shape

I did not live to see the World Wide Web transform society in the 1990s. I died in 1990, just as that new era was beginning. But many of the ideas I had spent decades discussing were becoming real. Computers were becoming personal, interactive, connected, and increasingly useful as communication devices. Networks were no longer merely experiments linking machines; they were beginning to connect people, ideas, libraries, institutions, and communities.

An Idea Larger Than Any One Researcher

I never considered myself the inventor of the Internet. No one person deserves that title. My contribution was largely to ask what computers might become before the necessary technology fully existed, and then to help researchers pursue those possibilities. If I owe anyone an apology, perhaps it is for calling my vision a “Galactic Network.” The name may have sounded extravagant in 1962. Looking at the billions of interconnected devices that followed, I may actually have been thinking too small.

MIT Imagines a World of Connected Computers - Told by J.C.R. Licklider

At the beginning of the 1960s, computers were impressive machines, but they were remarkably lonely ones. Most operated as isolated systems. A researcher prepared instructions, submitted them to a computer, and often waited for the machine to return an answer. I thought that was an extraordinarily poor way to use such powerful equipment. Computers should not merely calculate answers after people had already formulated every question. They should participate in the process of thinking, communicating, and discovering.

Computers Should Work With Us

In 1960, I had written about what I called “man-computer symbiosis.” My argument was not that computers should replace human beings. I imagined a partnership. People would establish goals, formulate questions, recognize patterns, and exercise judgment. Computers would handle calculations, retrieve information, compare possibilities, and perform repetitive work at speeds no person could match. To make that partnership practical, however, people needed to interact directly with computers instead of waiting hours or days for batches of punched cards to be processed.

MIT Begins Thinking About Communication

MIT was becoming one of the great centers of this new approach to computing. Researchers were experimenting with time-sharing, interactive systems, graphical computing, communications, and ways of allowing multiple people to use computing resources. One particularly important researcher was Leonard Kleinrock. In July 1961, while at MIT, Kleinrock published theoretical work showing how data communications using queued, packet-like transmissions could operate efficiently. Mathematics was beginning to demonstrate that computers did not necessarily need the kind of dedicated circuits traditionally used by telephone networks.

Why Should Every Computer Be an Island?

Once people could interact directly with computers, another question became difficult to ignore. Why should every computer remain isolated? A university might possess valuable programs and information on one machine while researchers hundreds of miles away had completely different resources on another. Connecting them could mean that the usefulness of each computer would no longer be confined to the building containing it. The real opportunity was not simply faster calculation. It was access—allowing people to reach information, programs, and eventually one another through machines.

The Intergalactic Computer Network

During 1962, I began circulating memoranda discussing what I playfully called an “Intergalactic Computer Network.” The name was intentionally grand, but the concept was serious. I imagined computers connected through a network in which people could access programs and data located at other sites. Where the information physically resided would become much less important. A researcher sitting at one terminal might reach resources stored on another machine somewhere else in the country—or eventually around the world. The Internet Society has described this vision as remarkably similar in spirit to the Internet that later emerged.

The Machine Was Becoming a Communication Device

This represented an important change in thinking. A computer had traditionally been understood as a machine to which a person submitted calculations. I was increasingly interested in the computer as something else: a communication medium. Once computers were interconnected, people could potentially exchange information through them, collaborate across long distances, and share resources that had previously been trapped inside individual machines. The computer would no longer simply help one person calculate. It could help communities of people think together.

Taking the Vision to ARPA

In October 1962, I became the first head of ARPA's computer research program, later known as the Information Processing Techniques Office. I did not arrive carrying blueprints for ARPANET—that network would come later. What I brought was the conviction that interactive computing and interconnected computers deserved serious investment. I began supporting researchers who were exploring time-sharing, graphics, artificial intelligence, and computer communications, while encouraging people such as Ivan Sutherland, Bob Taylor, and eventually Larry Roberts to continue pushing these ideas forward.

Imagining the Internet Before It Existed

In 1961 and 1962, there was still no Internet, no Web browser, no Google search bar, and no laptop waiting on a kitchen table. Yet many of the questions that would shape that future were already being asked. Kleinrock was demonstrating mathematically how digital communications networks could operate. MIT researchers were developing interactive computing. I was asking what might happen if computers everywhere could be connected and their information made accessible to people wherever they happened to be. We did not yet possess all the technology necessary to build that world. But before engineers could construct such a network, somebody first had to imagine why anyone would want one.

ARPA Turns the Ideas into an Experimental Computer Network - Told by Licklider

In 1962, I arrived at the Advanced Research Projects Agency with a conviction that computers could become much more than isolated calculating machines. I imagined people working interactively with computers and eventually using connected machines to reach information and programs stored elsewhere. An idea, however, is only the beginning. ARPA possessed something equally important: the ability to fund talented researchers at different universities and laboratories and give them room to attempt things that conventional organizations might consider too experimental. That environment would turn the dream of connected computing into machines that actually communicated.

Building a Community Before Building a Network

As director of ARPA's new Information Processing Techniques Office, I began supporting research at places such as MIT, Stanford, Carnegie Tech, UCLA, and other institutions. One of our major projects was MIT's Project MAC, which explored time-sharing and interactive computing. Instead of treating a computer as a machine that belonged to one user or one calculation at a time, time-sharing allowed many people sitting at terminals to interact with a powerful computer seemingly at once. Something unexpected happened: users did not simply share the machine. They began forming an online community, exchanging programs, messages, and information. We were discovering that when people shared computers, they also began using computers to communicate with one another.

The Problem of Too Many Computers

There was an inconvenience hidden inside ARPA's success. We were paying for different computers at different research centers, and those machines often could not communicate with one another. A researcher who wanted to use several systems might need several terminals, each connected to a different computer. To me, this represented exactly the problem I had been talking about. Why should a person care where a program or piece of information physically resided? In an April 1963 memorandum to our research community, I described a future in which connected computers could become an essential medium for informational interaction among people and institutions. The machines needed to stop behaving like isolated islands.

Others Take the Idea Forward

I left ARPA in 1964, so I cannot honestly tell you that I personally built ARPANET. I did not. The people who followed me transformed the ideas into an engineering project. Ivan Sutherland succeeded me, and then Bob Taylor became director of the Information Processing Techniques Office. Taylor strongly supported connecting ARPA's scattered research computers. Larry Roberts, who had already participated in experiments connecting distant computers, was recruited to lead the networking effort. By 1967, plans for what became ARPANET were taking recognizable form. What had once been an argument about why computers should communicate was becoming a question of exactly how engineers would make them do it.

A New Kind of Network

The engineers faced a difficult problem. The computers at these research centers were not identical. They had different manufacturers, operating systems, capabilities, and purposes. ARPA therefore proposed placing specialized machines called Interface Message Processors, or IMPs, between the host computers and the communications network. Rather than requiring every large computer to understand the entire network, the IMP would handle much of the task of moving information from place to place. In 1968, ARPA awarded Bolt Beranek and Newman, or BBN, the contract to build them. Frank Heart led the BBN team, with researchers including Bob Kahn helping design the system.

The First Machines Are Connected

In September 1969, BBN delivered the first IMP to Leonard Kleinrock's laboratory at UCLA. A second went to Douglas Engelbart's group at the Stanford Research Institute. Teams at each location had to make their enormous host computers communicate through these new machines. Meanwhile, young researchers including Steve Crocker, Vint Cerf, and Jon Postel were working on the rules that computers would use to communicate. Crocker began circulating documents called Requests for Comments, or RFCs—a wonderfully informal name for documents that would eventually help define Internet standards for decades.

The Night the Network Spoke

On October 29, 1969, programmer Charley Kline sat at UCLA attempting to log into the computer at SRI, roughly 350 miles away. He typed an “L.” It arrived. He typed an “O.” It arrived. Then he typed “G”—and the receiving computer crashed. The first message sent between the two ARPANET hosts therefore became simply “LO.” The problem was quickly corrected, and the complete login soon succeeded. By the end of 1969, UCLA, SRI, UC Santa Barbara, and the University of Utah formed the original four-node ARPANET. An idea that only a few years earlier existed in memoranda, discussions, mathematics, and research proposals had become a working computer network.

From an Idea to Something Real

ARPANET was not yet the Internet. It was an experimental network, and many enormous breakthroughs still lay ahead. But something fundamental had changed between 1962 and 1969. We had begun the decade with expensive computers operating largely as separate machines. We ended it with different computers at distant research institutions exchanging information across a common network. I had spent my time at ARPA encouraging researchers to imagine computers as tools for communication and cooperation. Taylor, Roberts, Kleinrock, Heart, Engelbart, Crocker, Cerf, Postel, Kahn, and many others carried that vision into hardware and software. The important achievement was never that one person invented the network. It was that enough people finally believed connected computers were worth building—and then figured out how to make them talk.

My Name is David Boggs: Helping Build Ethernet

I was born in Washington, D.C., in 1950, and I became fascinated with electronics and radio long before computer networks became part of everyday life. I was an amateur radio operator, and that experience taught me an important lesson: communication did not require a perfectly controlled path. Signals could interfere, disappear, and return, yet a properly designed system could still deliver information reliably. That way of thinking would become surprisingly useful when I began working with computers.

Arriving at Xerox PARC

In the early 1970s, I was a graduate student at Stanford when I began spending time at Xerox's Palo Alto Research Center, better known as PARC. It was an unusual place. Researchers were building the Alto, an early personal computer, along with graphical displays, laser printers, and other technologies that seemed far ahead of what most offices were using. But there was an obvious engineering problem: if every person had a computer, those computers needed an efficient way to communicate with one another and share expensive devices such as printers.

Meeting Bob Metcalfe

Robert Metcalfe was working on exactly that problem. He had been studying computer networking and was familiar with ALOHAnet, a wireless network developed at the University of Hawaii. In 1973, he proposed a new local network that would use a shared coaxial cable instead of radio. I joined him in turning that concept into working hardware. Metcalfe developed much of the architecture, while I concentrated heavily on the electronics and implementation. We had a concept, but an inventor quickly discovers that a diagram and a functioning machine are two very different things.

Making Ethernet Work

We called the system Ethernet. Computers attached to the same cable could listen before transmitting. If two machines transmitted at the same time, they could detect the collision, stop, wait, and try again. It was a remarkably simple idea compared with systems that tried to carefully schedule every transmission. My background in radio helped me understand that a shared and somewhat unpredictable communications medium could still be made reliable. By the end of 1973, we had an experimental Ethernet connecting computers at PARC.

Connecting the Office of the Future

Ethernet allowed Xerox Alto computers to communicate with one another, reach servers, and share laser printers. That might sound ordinary today, but in the 1970s most people had never touched a personal computer, much less used one connected to a local network. PARC was effectively experimenting with a networked office years before such offices became commonplace. Ethernet eventually faced competitors such as Token Ring and ARCNET, but its simplicity, speed, and ability to evolve helped it become the dominant local-area networking technology.

From Local Networks to a Connected World

Our work did not stop at connecting computers inside one building. Researchers at PARC also experimented with connecting Ethernet networks to ARPANET and other systems using the PARC Universal Packet architecture. The larger challenge was becoming clear: individual networks were useful, but connecting different networks together could create something much more powerful. That same basic principle of linking local networks into larger networks became fundamental to the Internet that followed.

Watching Ethernet Become Invisible

I continued working in computer networking and research long after those early experiments. Over time, Ethernet became so successful that people stopped thinking about it. Cables ran through office walls, schools, laboratories, factories, data centers, and eventually homes. Faster versions replaced the original experimental system, but the underlying idea continued to evolve rather than disappear. A technology can reach a peculiar kind of success when people no longer notice that they are using it.

Building What Others Had Only Imagined

I died in 2022, nearly fifty years after those first Ethernet experiments. I never believed that Ethernet appeared from nowhere. We learned from earlier networking work, particularly ALOHAnet, ARPANET, and researchers who had been experimenting with packet communications for years. Our contribution was to solve a particular problem: how to make nearby computers communicate quickly, cheaply, and reliably. If I owe an apology for anything, perhaps it is that inventors sometimes make their work sound simpler afterward than it really was. The principle may have been straightforward. Making the hardware behave was another matter entirely.

Xerox PARC Ethernet: Connecting the Computers in the Building - Told by Boggs

By 1973, researchers had already proven that computers hundreds of miles apart could communicate through networks such as ARPANET. At Xerox's Palo Alto Research Center, however, we had a different problem. We were building an office filled with personal computers. The Xerox Alto was appearing on researchers' desks, laser printers were being developed, and servers could hold information everyone wanted to use. We did not need to connect California to Massachusetts. We needed to connect the computer sitting in one office to another machine fifty feet down the hallway—and we needed to do it quickly.

An Office Filled With Computers

You have to remember how unusual PARC was. Most organizations in 1973 did not have a computer sitting on every employee's desk. Computers were normally large, expensive machines shared by many users. At PARC, researchers such as Chuck Thacker, Butler Lampson, Alan Kay, and others were imagining something different: an office in which individuals had their own interactive computers. But once you put hundreds of computers in a building, isolation becomes ridiculous. A researcher should not have to carry information physically from one machine to another, and we certainly could not afford a separate laser printer for every desk. Our computers needed a common communications system.

Bob Metcalfe Has an Idea

Bob Metcalfe took on the networking problem. On May 22, 1973, he wrote a memo describing what he called the “Ether Network.” His thinking drew partly from ARPANET and particularly from ALOHAnet, an experimental radio network developed at the University of Hawaii. ALOHAnet allowed multiple computers to share the same radio channel rather than assigning every machine its own dedicated connection. Bob realized that a similar principle could work inside PARC—but instead of sending signals through the air, we could send them through a shared cable. I joined him because turning that idea into working electronics was exactly the sort of problem I enjoyed.

One Cable, Many Computers

Our network used coaxial cable running through the building. Every connected computer shared that communication channel. When a machine had information to transmit, it listened to determine whether another computer was already using the cable. If the line was clear, it transmitted. That sounds simple until two machines decide to transmit almost simultaneously. Then their signals collide. Rather than attempting to eliminate collisions entirely, Ethernet was designed to detect them. The computers could stop transmitting, wait for slightly different periods of time, and try again. It was a practical solution to a messy problem: allow the machines to compete for access, but give them rules for recovering when they interfered with one another.

Radio Had Taught Me Something Useful

My background as an amateur radio operator helped this idea make sense to me. Radio operators already understood that many people could share a communications medium. You listened before transmitting, interference sometimes occurred, and a successful system needed procedures for recovering from it. Ethernet treated the cable somewhat like a miniature radio environment. In fact, the name itself referred to the old scientific concept of the “ether,” an invisible medium once believed to carry electromagnetic waves. The ether of our network was not imaginary—it was a real cable—but the name captured the idea of many machines communicating through a common medium.

Making the Hardware Behave

This was where invention stopped being a memo and became engineering. Bob and I needed hardware that could transmit information quickly, detect collisions, recognize addresses, and communicate reliably enough that researchers would actually trust it. The network also had to be fast enough to serve Xerox's experimental laser printers, which required moving substantial amounts of data. Our original Ethernet operated at roughly 2.94 megabits per second—extraordinarily fast for a local computer network of that period. By November 1973, our experimental system was beginning to operate inside PARC, connecting the new Alto computers and demonstrating that the basic concept worked.

The Office Begins to Look Like the Future

Once computers could communicate through Ethernet, something larger happened. Alto users could reach shared servers and laser printers instead of treating each computer as an isolated machine. Electronic communication also became part of everyday life inside PARC. Researchers exchanged messages, shared files, collaborated on projects, and formed electronic discussion groups. If you had walked through PARC in the mid-1970s, you would have seen something that looked remarkably familiar to someone from decades later: people sitting at personal computers connected through a local network, sending messages and accessing shared resources. The rest of the world simply had not caught up yet.

A Local Network Becomes Part of a Global One

Ethernet did not create the Internet, and we never claimed that it did. ARPANET already existed, packet networking had been developed by earlier researchers, and Bob Kahn and Vint Cerf were beginning the work that produced TCP/IP. Our contribution solved another essential problem: how to connect large numbers of computers efficiently within a local area. PARC was even experimenting in 1973 with connecting Ethernet networks to other networks using its PARC Universal Packet system. Eventually technologies such as TCP/IP would make it possible for local networks like Ethernet to become pieces of a much larger Internet.

The Cable That Refused to Disappear

Ethernet eventually faced formidable competitors, including IBM's Token Ring and other local-networking systems. Yet Ethernet continued evolving. It became faster, cheaper, standardized, and increasingly widespread until it became the dominant form of wired local networking. The system Bob Metcalfe and I worked on in 1973 was primitive compared with today's networks, but the underlying idea proved remarkably durable. We were trying to connect a building full of experimental computers and printers. We could not have known just how many homes, schools, businesses, factories, and data centers would eventually depend upon descendants of that work. Sometimes an invention begins with a grand attempt to change the world. Ethernet began with a more practical question: we have all these computers—now how do we get them to talk to each other?

From Separate Networks to “The Internet”: TCP/IP - Told by J.C.R. Licklider

By the early 1970s, the problem had changed. We had already learned how to make computers communicate across a network. ARPANET was working, packet radio systems were being tested, satellite networks were emerging, and local networks such as Ethernet were appearing. But these networks were different from one another. They used different equipment, different rules, and different ways of moving information. The next great challenge was not simply building another network. It was finding a way to make all of these networks communicate as though they were parts of one larger system. That idea would become the Internet.

Bob Kahn Takes on the Problem

In 1973, Bob Kahn was working at DARPA and began concentrating on what he called internetworking. He wanted ARPANET to communicate with newer packet-radio and satellite networks without forcing every network to be rebuilt in the same way. That was a crucial idea. Each network should be allowed to operate according to its own internal design while still communicating with the others. Kahn established several principles: networks should remain independent, packets should be delivered on a best-effort basis, special gateways should connect the networks, and there should be no single global controller directing the entire system.

Vint Cerf Helps Create a Common Language

Kahn needed someone who understood the software already being used on ARPANET, so in 1973 he asked Vint Cerf at Stanford to help design the new system. Cerf had worked on ARPANET's earlier Network Control Protocol and understood the difficulties of getting different computers to communicate. Together, Kahn and Cerf developed a new protocol that could move information not merely between computers on one network, but across entirely different networks. They presented an early version in 1973 and published their influential paper, “A Protocol for Packet Network Intercommunication,” in 1974.

A Network of Networks

This was the conceptual leap. A computer sending information would not need to know every detail about all the networks between itself and its destination. Packets could travel through gateways from one network into another. An ARPANET computer might send information through a packet-radio network or a satellite connection and eventually reach another computer somewhere else. The individual networks remained separate internally, but to the user they could increasingly behave like one interconnected system. The term “Internet” grew naturally from this idea of internetworking—a network made from other networks.

Proving That It Could Actually Work

A design on paper was not enough. DARPA funded implementations of the new protocol at Stanford, BBN, and University College London. One of the most dramatic demonstrations occurred in 1977. A computer in a moving packet-radio van in California transmitted information through the Packet Radio Network, into ARPANET, across the Atlantic through the SATNET satellite network, and eventually back through ARPANET. Three very different networks had successfully communicated using a common internetworking protocol. It was a powerful demonstration that an Internet could actually function.

TCP Becomes TCP/IP

Originally, the system was largely described as TCP, the Transmission Control Protocol. As researchers experimented with different applications, however, they discovered that one protocol was being asked to do too many things. Some applications needed reliable delivery and correction when packets were lost. Others, such as experimental voice communications, sometimes needed speed more than perfect retransmission. The design was therefore divided. Internet Protocol, or IP, would handle addressing and moving individual packets between networks. Transmission Control Protocol, or TCP, would handle reliable delivery, ordering, and recovery when necessary. Together they became known as TCP/IP.

The Internet Starts Spreading

During the late 1970s and early 1980s, TCP/IP was tested across more machines and networks. Researchers also helped place TCP/IP into the Berkeley version of Unix, making the protocols available to a much wider university and research community. At the same time, Ethernet networks were spreading inside laboratories and institutions. TCP/IP could connect these local networks to other networks farther away. The pieces were fitting together: Ethernet could connect machines inside a building, while TCP/IP could help connect that building's network to networks elsewhere.

January 1, 1983

Eventually ARPANET had to make a choice. Its older Network Control Protocol could not support the expanding world of interconnected networks nearly as effectively as TCP/IP. The Department of Defense established TCP/IP as its standard and ordered ARPANET hosts to make the transition. On January 1, 1983, ARPANET officially switched from NCP to TCP/IP. Computers that failed to convert could no longer communicate normally with the network. This “flag day” became one of the most important milestones in Internet history because TCP/IP was now the common language connecting networks together.

The Vision Becomes an Internet

I had imagined interconnected computers back in the early 1960s, but what emerged by 1983 was far more sophisticated than simply linking one machine to another. Kahn, Cerf, Jon Postel, Ray Tomlinson, Peter Kirstein, and many other researchers had created something more flexible: a system in which entirely different networks could join together without surrendering their individuality. There was no single central machine controlling everything. New networks could be added as the system expanded. That principle is one reason the Internet could eventually grow from a research experiment into a global communications system. We had stopped thinking only about connecting computers. We had learned how to connect networks themselves.

The World Wide Web Changes Everything — 1989–1993 - Told by David Boggs

By the end of the 1980s, computer networking had already traveled an extraordinary distance. ARPANET had demonstrated long-distance packet networking, Ethernet had connected computers inside buildings, TCP/IP allowed different networks to communicate, and universities and research centers around the world were joining the Internet. Yet using the Internet was still not something most ordinary people found simple. You needed to know where information was located and often which program or command would retrieve it. Then a researcher working in Switzerland proposed something that would make all of those connected machines vastly easier to explore.

Tim Berners-Lee Sees a Different Problem

In 1989, British computer scientist Tim Berners-Lee was working at CERN, the enormous European physics laboratory near Geneva. CERN's scientists came from universities and laboratories around the world, bringing different computers, software, and methods of organizing information. Berners-Lee saw researchers struggling to keep track of documents, experiments, people, and projects scattered across incompatible systems. His proposal was deceptively simple: use hypertext to connect documents across a computer network, allowing a reader to follow links from one piece of information to another. CERN needed better information sharing. Berners-Lee's solution would eventually change how the world used the Internet.

The Internet and the Web Were Not the Same Thing

This distinction is important. Berners-Lee did not invent the Internet. The Internet already existed. Ethernet, TCP/IP, routers, servers, cables, and other technologies provided the roads over which digital information could travel. What Berners-Lee created was a new system for organizing and reaching information traveling across those roads. Think of the Internet as the transportation network and the World Wide Web as an enormous collection of destinations connected by signs that allow you to move easily from one to another. The Web ran on top of the Internet and made the network far more useful to ordinary people.

Three Ideas Come Together

During 1990, Berners-Lee transformed his proposal into a working system. He created HyperText Markup Language, or HTML, for constructing documents; Hypertext Transfer Protocol, or HTTP, for requesting and delivering those documents; and Uniform Resource Locators, or URLs, for giving resources identifiable locations. Working with Robert Cailliau and with support inside CERN, he built the first Web server and the first browser-editor on a NeXT computer. By Christmas 1990, the basic pieces of the World Wide Web were functioning together.

The First Website Appears

The first website was not glamorous. It explained what the World Wide Web was, how to obtain browser software, and how people could create Web servers of their own. But that simplicity was part of the brilliance. A document could contain a highlighted link to another document. Select the link, and the computer would retrieve the new information. That document could contain additional links leading somewhere else. Suddenly information no longer had to be read as a single isolated file. It could become part of a growing web of connected knowledge. CERN identifies Berners-Lee's server as the home of the world's first website.

Opening the Web Beyond CERN

The idea began spreading outside CERN. Berners-Lee encouraged other researchers to install servers and create pages, while new browsers allowed people using different types of computers to participate. A line-mode browser released in 1992 helped make Web access possible on a broader range of machines. The important principle was openness: people did not need permission from one central company to create another Web server or connect another collection of pages. The system could grow wherever people chose to add information.

April 30, 1993: CERN Opens the Door

One decision in 1993 proved enormously important. On April 30, CERN announced that key World Wide Web software would be made available on a royalty-free basis, allowing people to use, copy, modify, and distribute it without paying CERN licensing fees. That helped ensure that the Web would not remain a proprietary system controlled by one organization. Developers, universities, companies, and individuals could build upon it. By the end of 1993, CERN reported more than 500 known Web servers. Something that had begun as a solution for physicists was escaping the laboratory.

Then Came Mosaic

Another development in 1993 made the transformation impossible to ignore. At the National Center for Supercomputing Applications at the University of Illinois, Marc Andreessen and Eric Bina developed Mosaic, a graphical Web browser that was comparatively easy to install and use. It displayed text and images together and was eventually made available for major computer platforms. People no longer had to think like computer-network specialists merely to explore information. They could point, click, follow links, see photographs, and move through websites visually. Mosaic helped push the Web from a research tool toward mass use.

We Had Built the Roads—Now People Had Somewhere to Go

From my perspective as one of Ethernet's developers, the Web demonstrated why the networking work of the previous twenty years mattered so much. Ethernet could connect computers inside a building. TCP/IP could connect those networks across enormous distances. But Berners-Lee gave people an intuitive way to move through the information stored on them. By 1993, the Internet was no longer merely an impressive engineering system connecting computers. The World Wide Web was turning it into a place people could explore. Once ordinary users discovered that clicking one link could carry them to information stored on another computer somewhere across the world, there was very little chance of putting that idea back into the laboratory.

My Name is Jim Kimsey: Building America Online

I was born in Washington, D.C., in 1939, and I did not begin my career expecting to build one of America's most recognizable technology companies. I grew up without much money, attended West Point, and graduated in 1962. The Army taught me lessons that stayed with me throughout my business career: understand the people around you, make decisions when you do not have perfect information, and don't expect a plan to survive unchanged once the real work begins.

Lessons From Vietnam

I served as an Army Ranger and completed two tours in Vietnam. Commanding soldiers was considerably different from running a company, but leadership has certain similarities wherever you find it. You have to recognize talent, give people responsibility, and make decisions when circumstances change. I also became deeply connected to the Vietnamese people, helping complete an orphanage near Duc Pho and continuing to support humanitarian efforts there long after the war ended. Those experiences affected how I looked at responsibility for the rest of my life.

Becoming an Entrepreneur

When I returned to Washington, I tried the investment business and eventually put about $2,000 into a building and entered the restaurant business almost by accident. I went on to operate places including the Exchange, Bullfeathers, and the Madhatter. They became successful, but eventually I grew bored. Building something interested me more than maintaining it. Through a West Point connection, I was introduced to a struggling technology company called Control Video Corporation. The company was trying to deliver games electronically to home computers. It had serious problems, but I thought the underlying idea had potential.

A Company That Almost Didn't Survive

Steve Case and I joined Control Video while it was struggling to stay alive. I became its chairman and chief executive, while Steve concentrated increasingly on marketing and the consumer. We restructured the operation and, in 1985, created Quantum Computer Services. Our Q-Link service for Commodore computers provided games, information, electronic communication, and other services. When Q-Link launched on November 1, 1985, only 24 people logged on during its first fourteen hours. That was hardly evidence that we were about to change the world. But entrepreneurs learn not to confuse a small beginning with a bad idea.

Building America Online

We kept adapting. Quantum expanded beyond Commodore computers and developed services for other personal computers. Eventually we needed an identity of our own. In 1989, the name America Online emerged. The idea was straightforward but ambitious: online services shouldn't belong exclusively to computer experts. Ordinary families should be able to turn on a computer, connect through a modem, send messages, read information, play games, and communicate with other people without understanding what was happening behind the screen.

Knowing When Someone Else Should Lead

One of my most important business decisions was recognizing that Steve Case understood where the consumer market was heading. My role was often to raise money, work with investors, recruit credible people, negotiate deals, and keep the organization alive long enough for its ideas to succeed. Steve increasingly drove strategy and marketing. Eventually I moved aside and let him run the company. Years later I joked that knowing when to get out of the way made me look like a genius. Leadership isn't proving that you're the smartest person in the room. Sometimes it means recognizing who should be running the room.

When America Went Online

By the middle of the 1990s, AOL had millions of subscribers. People who had never considered themselves computer enthusiasts were choosing screen names, sending email, joining chat rooms, reading news, and meeting people hundreds or thousands of miles away. In 1995, AOL already had more than four million users. What had once looked like a strange little business involving computers and telephone lines was becoming part of American daily life.

Build It, Then Let It Grow

I stepped down as AOL's chairman in 1995 and devoted much of my later life to philanthropy, education, veterans, and civic projects. I died in 2016, long after the Internet had moved far beyond anything we imagined during those difficult early days at Control Video. I cannot claim to have invented the Internet; scientists and engineers had been constructing its foundations for decades before I arrived. What we helped do was make the online world understandable and useful to ordinary people. If I owe an apology for anything, perhaps it is that we made logging on seem so simple that an entire generation forgot how extraordinary it once was to connect one household computer to the rest of the world.

Mosaic Makes the Internet Something Ordinary People Can See - Told by Kimsey

By 1993, the Internet already existed, the World Wide Web had been created, and researchers had been exchanging information electronically for years. But from a businessman's point of view, there was still a serious problem: most ordinary people had no idea why they should care. If a technology requires you to understand commands, file structures, protocols, and unfamiliar software before you can enjoy it, you have dramatically reduced your potential audience. Then Mosaic arrived and changed what people thought the Internet could look like.

Two Young Developers at NCSA

At the National Center for Supercomputing Applications at the University of Illinois, Marc Andreessen and Eric Bina began developing Mosaic. Other Web browsers already existed, including Tim Berners-Lee's original browser and programs such as ViolaWWW. Mosaic's importance was not that it invented browsing. Its importance was that Andreessen, Bina, and the NCSA team concentrated on making browsing easier, more reliable, and more appealing to people who were not computer specialists. NCSA released the first Unix version in 1993, and versions for Macintosh and Microsoft Windows followed.

The Web Suddenly Had a Face

One of Mosaic's most important features was the way it presented graphics. Earlier browsers often handled images separately from the text surrounding them. Mosaic helped popularize pages in which pictures appeared directly alongside words. That seems painfully ordinary now, but it changed the experience completely. A Web page could begin to look less like a technical document and more like a magazine, catalog, newspaper, or illustrated book. People did not simply retrieve information anymore. They could look around.

Point, Click, Go Somewhere Else

Mosaic also made navigation much easier to understand. Buttons for moving backward, forward, or home, bookmarks for saving interesting locations, and clickable links helped turn the Web into something people could explore without first becoming network engineers. You saw something interesting, clicked it, and another page appeared. Then you clicked again. From a consumer-business perspective, that was enormously important. Every layer of difficulty you remove gives another group of people permission to participate.

The Downloads Start Pouring In

The response came quickly. By December 1993, more than 5,000 copies of Mosaic were reportedly being downloaded each month, and NCSA was receiving an extraordinary volume of inquiries about the software. The browser became prominent enough to receive national press attention, while NCSA increasingly devoted resources to supporting it. What had begun inside a supercomputing center was attracting people far beyond the scientific community.

Why This Mattered to Companies Like Ours

At America Online, we had been working for years to make online services understandable to ordinary consumers. Our users could already send messages, join discussions, read information, and play games through an online service designed around ease of use. Mosaic showed that the open Internet was beginning to solve the same consumer problem. The Internet no longer had to look like an intimidating collection of commands and addresses. The Web could become visual, approachable, and increasingly useful to someone sitting at home with a personal computer.

From Research Tool to Mass Medium

During 1994, the momentum accelerated. Mosaic was available across major computing platforms, more websites were appearing, and developers were realizing that the browser could become the main doorway into the Web. Andreessen left NCSA that year and soon helped launch the company that produced Netscape Navigator, but Mosaic had already demonstrated the larger principle. A graphical browser could transform a technical network into a mass communications medium. The Computer History Museum describes Mosaic as helping move the Web from a research project toward widespread success.

The Internet Stops Looking Like a Laboratory

That was the real breakthrough. Mosaic did not invent the Internet, TCP/IP, or the World Wide Web. It made what those earlier inventors had built easier for ordinary people to see and understand. Once users could open a browser, view pictures, read text, click links, and move from one site to another, the Internet began feeling less like infrastructure and more like a destination. For businesses like AOL, that was both an opportunity and a warning. Millions of people were about to discover that going online did not have to be something reserved for computer experts. It could be something almost anyone might want to do.

America Dials In: AOL, Modems, Email, & Family Computer - Told by Jim Kimsey

By the middle of the 1990s, computers were beginning to move from offices and hobbyist rooms into ordinary American homes. The challenge was no longer simply building networks. It was convincing families that going online was worth the trouble. That meant making the experience simple enough that a parent, teenager, teacher, or first-time computer owner could connect without knowing anything about protocols, servers, or networking. At America Online, we believed that if we could remove enough of the difficulty, millions of people would discover that being online was useful, entertaining, and surprisingly social.

The Sound of Going Online

For most families, the journey began with a modem and a telephone line. You connected the modem to the computer, the modem dialed a local access number, and then came that unforgettable electronic sequence of tones, chirps, and static as two machines negotiated a connection. While the computer was online, the household telephone line was usually busy. Connections were slow by modern standards, but to someone experiencing it for the first time, the fact that a computer in the living room could communicate with machines hundreds or thousands of miles away seemed extraordinary.

Making the Online World Simple

AOL's real advantage was not that we had invented the Internet. We had not. Our job was to make the online experience approachable. Instead of dropping users into a confusing command line, we gave them menus, icons, categories, and recognizable destinations. They could read news, check weather, play games, visit discussion areas, send messages, or enter chat rooms. The computer stopped looking like a machine that demanded technical expertise and started behaving more like a doorway.

“You’ve Got Mail”

Email became one of the strongest reasons people returned. A person could write a message in the evening and someone across the country might read it minutes later. Families kept in touch, students communicated with friends, and businesses began realizing that electronic messages could move faster than letters or faxes. AOL's familiar “You've Got Mail” announcement became one of the defining sounds of the 1990s because it turned something technical into something personal: someone had written to you.

The Disks Were Everywhere

We wanted people to try the service, so AOL distributed enormous numbers of free-trial disks through magazines, stores, direct mail, and promotional partnerships. You could find them almost everywhere. From a business standpoint, the logic was simple: every household with a computer was a potential customer, but many people needed a reason to take the first step. The disks lowered that barrier. Install the software, plug in the modem, dial the number, and suddenly the online world was sitting in your home.

Chat Rooms Change the Meaning of a Computer

Perhaps the most surprising development was how quickly people used computers to talk to other people. Chat rooms allowed strangers with common interests to meet in real time. There were rooms for hobbies, sports, politics, parenting, entertainment, and almost anything else users wanted to discuss. Screen names gave people a kind of digital identity. For better and for worse, users discovered that they could meet people they never would have encountered in their own neighborhoods. The computer was becoming a social machine.

AOL Connects to the Open Internet

During this period, AOL also moved beyond being a closed online service. As the World Wide Web expanded, users increasingly expected access to the larger Internet. AOL added Internet capabilities and Web access, allowing members to move from AOL's own content into websites located elsewhere. That was a major shift. The future was clearly not going to belong to one private online service. It would belong to millions of interconnected destinations.

Millions of Americans Come Online

The growth was tremendous. AOL passed one million members in 1994, and by 1996 its membership had climbed into the millions. Personal computers were becoming more common, modem prices were falling, Windows-based machines were easier to use, and the Internet was becoming part of everyday conversation. For many Americans, AOL was their first experience with email, chat rooms, online communities, and the Web. Once people learned what was available, many of them wanted more.

The Family Computer Becomes a Window to the World

By 1996, the family computer was no longer just a machine for homework, spreadsheets, or games. It was becoming a communications center. A child could research a school project, a parent could send email, a family could read news from another state, and friends could talk without making a long-distance telephone call. The technology underneath all of this had taken decades to develop, but the revolution became real when ordinary families began using it every day. The Internet had finally left the laboratory, crossed the office, entered the home, and become part of American life.

Browsers, Search Engines, Chat Rooms, and Online Communities - Told by Kimsey

Between 1994 and 1998, going online changed from something people experimented with into something millions began building into their everyday routines. The technology underneath the Internet had existed for years, but now people needed practical ways to move through it, find what they wanted, and communicate with one another. Browsers gave them a doorway, search engines gave them a map, and chat rooms gave them something perhaps even more important: other people.

Netscape Opens the Web

In 1994, Marc Andreessen and Jim Clark helped launch Netscape Communications, and Netscape Navigator quickly became one of the most popular Web browsers. It built on ideas demonstrated by Mosaic but made browsing faster and more practical for a rapidly growing audience. Suddenly people could open a program, type an address, and travel from one website to another by clicking links. The browser was becoming one of the most important pieces of software on the personal computer because it gave ordinary users a simple way into an Internet that had once seemed designed mainly for researchers and computer enthusiasts.

Microsoft Joins the Browser Battle

Microsoft quickly recognized what was happening. Internet Explorer appeared in 1995 and became increasingly integrated with Windows. Netscape and Microsoft then entered what became known as the browser wars, each trying to add features, improve speed, and attract more users. From the consumer's point of view, all that competition helped make the Web easier to use. Browsers improved rapidly because companies understood that whoever controlled the doorway to the Internet might influence how millions of people experienced it.

The Internet Becomes Too Large to Explore Alone

There was another problem. As websites multiplied, simply knowing that information existed somewhere on the Internet was not enough. People needed a way to find it. Yahoo!, founded by Jerry Yang and David Filo, began as a human-organized directory of websites. AltaVista, launched in 1995, allowed users to search enormous numbers of Web pages by typing words into a search box. Excite, Lycos, and other search services competed as well. For ordinary users, this changed everything. Instead of having to know the exact address of a website, you could begin with a question and let a search engine help you discover where the answer might be.

AOL Chat Rooms Fill With People

At AOL, we saw another side of the revolution. People were not logging on only because they wanted information. They wanted conversation. Our chat rooms allowed users to enter spaces devoted to music, sports, hobbies, parenting, politics, relationships, entertainment, and countless other subjects. Some people stayed for hours. They developed friendships with people they had never met face to face. Screen names became identities, and regular users began recognizing one another the way neighbors might recognize familiar faces at a local gathering place.

A New Kind of Community

This was something much larger than electronic mail. Online communities began developing their own customs, jokes, arguments, friendships, and personalities. Services such as AOL, bulletin-board systems, Usenet discussion groups, and Web forums allowed people with unusual interests to find one another even when they lived hundreds of miles apart. A teenager interested in astronomy, a parent looking for advice, or a collector searching for someone with the same hobby could suddenly locate an entire community. Geography mattered less than it had before.

Instant Messaging Makes Conversation Immediate

In 1997, AOL Instant Messenger, better known as AIM, appeared. Email was useful, but instant messaging felt more like a conversation. Users could create buddy lists, see when friends were online, and send messages back and forth almost immediately. For young people especially, this became an important part of social life. The computer was no longer simply something you used. It became a place where your friends might already be waiting.

The Internet Develops a Different Personality

Of course, online communication also introduced problems. People could hide behind screen names, pretend to be someone else, spread rumors, behave badly, or communicate with strangers without understanding the risks. Parents and schools began confronting questions about online safety, privacy, and appropriate behavior that had barely existed a few years earlier. The same anonymity that allowed someone to speak freely could also make dishonest or harmful behavior easier.

The World Online Becomes a Place of Its Own

By 1998, the Internet no longer felt like a collection of distant computers connected through telephone lines. It increasingly felt like a place. Browsers allowed people to explore it. Search engines helped them find their way through it. Chat rooms and message boards allowed them to gather inside it. Instant messaging allowed friendships to continue there in real time. We had entered the online business believing people would pay to reach information and services. What became clear during the 1990s was that one of the Internet's most powerful attractions was much simpler: people wanted to reach other people.

Licklider’s “Galactic Network” Becomes Reality — 1997–1999 - Told by Licklider

In 1962, I wrote about what I called a “Galactic Network”—a future in which computers around the world would be interconnected and people could quickly reach information and programs regardless of where those resources were stored. I did not live to see the closing years of the 1990s; I died in 1990. But if you want to understand what happened between 1997 and 1999, compare that world with the one we could only imagine three decades earlier. The computers were connected. Millions of people were using them. And the network was becoming something even more important: a place where human beings communicated, learned, searched, created, and formed communities. (internetsociety.org)

The Computer Was No Longer Alone

By 1997, the personal computer was increasingly expected to communicate with other computers. A modem could connect a household computer through a telephone line to an Internet service provider and from there to machines around the world. Ethernet connected computers inside schools and businesses, while TCP/IP allowed those local networks to become part of the larger Internet. The physical machine sitting on a desk was only the beginning. Its real value increasingly came from everything—and everyone—it could reach.

Millions of Americans Enter the Network

The numbers reveal how quickly the transformation was occurring. According to the U.S. Census Bureau, about 43 million American adults were already using the Internet from home, school, or work in 1997, with roughly 28 million using it from home. Email and searching for information were among the most common home uses. Two years later, tens of millions more adults had Internet access. What had once been the territory of scientists and computer specialists was becoming part of ordinary household life. (census.gov) (census.gov)

Information Begins to Feel Immediate

This was very close to what I had imagined. A student could search for information stored on computers far beyond the school library. A researcher could exchange documents with colleagues in another country. Families could communicate through email without waiting for a letter to cross the continent. News organizations, universities, government agencies, museums, businesses, and individuals were creating websites that anyone with an Internet connection might visit. The location of information was becoming less important than the ability to find it.

Finding Something in an Enormous Web

The growing Web created a problem we had anticipated in principle: access is useful only if people can locate what they need. Search services such as Yahoo!, AltaVista, Lycos, and Excite helped users navigate an expanding ocean of webpages. Then, in 1998, Google appeared with a new approach to ranking search results according in part to relationships among webpages. The very fact that sophisticated search engines had become necessary demonstrated how quickly the Web had grown. The challenge was no longer finding enough information. It was finding useful information among an overwhelming amount of it. CERN's own history describes the Web as combining personal computers, networking, and hypertext into a global information system. (worldwideweb.cern.ch)

The Network Becomes Social

Perhaps the development I would have found most interesting was not technological at all. People were using connected computers to form relationships. Email, discussion groups, message boards, chat rooms, and instant messaging allowed communities to form without requiring everyone to occupy the same physical place. A student fascinated by astronomy could find other enthusiasts across the country. Researchers could collaborate without sharing an office. Families separated by thousands of miles could exchange messages in minutes. The computer was becoming what I had hoped it might become: not merely a calculating device, but a medium through which people could work and think together.

A Revolution With Unequal Access

Yet this new world did not reach everyone equally. Internet use in the late 1990s varied considerably by income, education, age, and other circumstances. Computers still cost money, Internet subscriptions cost money, and many households lacked either the equipment or connections necessary to participate. The emergence of this “digital divide” demonstrated an important principle: creating a technology is not the same thing as making its benefits universally available. A truly useful information network had to be judged not merely by how many machines it connected, but by how many people could actually use it. (census.gov)

The Galactic Network Was No Longer Science Fiction

By 1999, no single person could control or even fully comprehend everything happening across the Internet. That was precisely part of its significance. Thousands of independent networks, millions of computers, countless websites, and growing communities of users had become interconnected through common standards. Back in 1962, I imagined a globally connected collection of computers through which people could access information and programs from almost anywhere. The Internet Society later observed that this vision was remarkably similar in spirit to the Internet that actually developed. I wish I had lived long enough to watch those final years of the 1990s unfold. I suspect my greatest surprise would not have been that computers finally learned to communicate with one another. It would have been how quickly human beings made that network their own. (internetsociety.org)