The History of the Web
If I were to ask you if you like using the Web you would probably say yes. My guess is you would tell me you have been using the Web for years; that you are completely comfortable reading text, clicking on links, looking at pictures and videos, downloading files, listening to music, and so on. Like many people, you probably use the Web every day, perhaps many times a day. And yet—
What if I asked you to define the Web? Could you tell me what it is?
My guess is, probably not. In fact, to me, the most interesting characteristic of the Web is that it is almost impossible to define exactly.
Why is this? After all, it's not difficult to define the "Internet" (the large, international computer network based on the family of IP protocols) or "email system" (a service used to send and receive written messages asynchronously over a computer network). What is it about the Web that makes thinking about it so difficult?
The answer lies in the observation that the Web is much more than a single Internet-based service. It is an amalgam formed from our biological nature melding with vast computational resources. Indeed, when you look at the Web with the right sort of eyes, what you see is nothing less than a glimpse into the future, a time when human beings will have fulfilled their destiny as a species.
Obviously, these are strong words. However, the idea that the Web is an expression of our biological destiny is an important one to remember if we are to discuss the history of the Web. In fact, what we call the Web is actually the fourth significant attempt on the part of human beings to create a large-scale, network-mediated extension of our species.
Over the last century, mankind has developed communication systems that have had a significant effect on our society: modern newspapers, the telephone, radio, television, communication satellites, mobile phones, and so on. These systems have changed how we interact as a species by making the mass dissemination of information quick, inexpensive, and reliable. At the same time, these capabilities have increased the pace of events drastically. As a result, life is getting faster and faster and, with each passing year, we find it more and more difficult to keep up.
However, these communication systems I just mentioned, for all their power, have one important limitation: there is no easy way for large numbers of individuals to participate. Although we may read a daily newspaper, watch a TV program, or listen to the radio, there is no way for us to use the same media to respond to what we see or hear; nor is there a way for us to contribute information of our own.
The Internet, of course, has changed all of that. For the first time, mankind has a means of disseminating information quickly while, at the same time, allowing people to respond and to contribute. In addition, the Internet allows connections between any two people, or among any group of people.
As a species, it is inherent in our nature to connect to one another. Human beings are nothing if not social. We talk to one another endlessly; we argue, discuss, trade information, and solve problems. What the Internet has shown us is that our desire to connect is actually a need to connect. Moreover, it is a biological need, much stronger and richer than anyone would have believed only a few decades ago.
The Web is nothing less than an effort to connect the human species in a significant and powerful way.
As you and I consider these points, we will see the true nature of the Web begin to emerge. We will find that the Web is not merely another Internet service or another cool thing to do with a computer. It is nothing less than an effort to connect the human species in a significant and powerful way.
History, as you know, is written by the winners, and the Web has been a huge success. Thus, when we talk about the origins of the Internet, we tend to pay a lot of attention to the early proponents of the Web. In effect, they are invited to write history in their own way.
The truth is, however, that there were many experiments to connect the human species using the Internet. Most of these experiments contributed something, and four of them were fabulously successful. The Web, as important as it turned out to be, was merely the fourth successful experiment. Moreover, the Web was, very definitely, built on the technology of the other three.
So let's take a few moments to back up and take a look at the real history of the Web. We'll start our journey in 1979, when a few early pioneers created the first real attempt to connect human beings in a widespread manner.
In 1979, a worldwide network was still a dream. It was a time before PCs and Macs, when most computer people used a terminal connected to either a mainframe or to a minicomputer running VMS or Unix. There were a few small computer networks, but they were rudimentary and isolated from one another.
The most important network of the time was the ARPANET, one of the ancestors of the Internet. In 1979, however, maintaining a network was expensive and problematic, and even though the ARPANET was already ten years old, it connected only 61 computers. To put this in perspective, as I write this, the Internet has well over 625 million computers, which makes it more than 10 million times larger than it was in 1979.
The first network email message had been sent eight years earlier (in 1971, by Ray Tomlinson at the technology company BBN). Still, in 1979, most ARPANET users found electronic mail new and exciting. And since email made it easy to send and receive messages quickly, humans beings had now reached a milestone: for the first time in history, people could use the written word to collaborate effectively at a distance.
Thus it came to pass that, as the decade ended, the biological urge to think and work collectively asserted itself in an unexpected way: a few of the early email users set up the world's first mailing list.
At the time, the ARPANET was supposed to be used only for research purposes. However, the very first mailing list had nothing to do with research, it had to do with a group of human beings needing talk to one another about something that was important to them. In fact, the first-ever network mailing list was created to discuss science fiction. (The list was called "sf-lovers".)
The invention of the mailing list meant that it was now possible for a group of people, spread around the world, to have discussions and work together asynchronously in something approaching real time. Within a few years, many mailing lists were created and, today, there are more lists than can be counted. In hindsight, it is obvious that the invention of the mailing list was the first successful computer-mediated implementation of a biological imperative: a need that compels us to connect to one another, no matter what the distance.
And yet— something was missing. Mailing lists were fine, but they were not enough: it was just too difficult to carry on public discussions. For one thing, you could only read messages if you subscribed to a list, and there was no way to search for mailing lists that might be of interest to you.
The second attempt to use the Internet to connect mankind was also based on a technology that was developed in 1979. In that year, two Duke University graduate students, Jim Ellis and Tom Truscott, developed a way to send news and announcements between two universities in North Carolina (University of North Carolina and Duke University). Within a short time, the system, called Usenet ("Users Network"), had spread to other schools, where it was developed into a sophisticated system of discussion groups.
As a tool to connect large numbers of individuals, Usenet worked much better than mailing lists. In fact, during the 1980s, the Internet itself was developed by people who used Usenet for their discussions and for distributing software. For many years, Usenet was the principal Internet vehicle for public discussions, and one of the two services used to distribute software (the other being "anonymous FTP"). Indeed, much of the software that ran the Internet was announced and distributed via Usenet. Without email and mailing lists, there would have been no Usenet. However, without Usenet, there would have been no Internet as we know it.
At first, Usenet supported only a handful of discussion groups, called "newsgroups" for historical reasons. Within a few years, however, the number of newsgroups had increased enormously, and it was not long before people around the world were talking, debating, collaborating, and sharing. Today, there are tens of thousands of Usenet discussion groups — still called newsgroups — covering every topic you can imagine (and a great many you can't imagine).
Compared to mailing lists, Usenet provided a much better means of connecting large numbers of people. Nevertheless, there were important limitations. Disk storage was expensive, and Usenet articles were deleted automatically after a short time, which made it difficult to accumulate information on a permanent basis. As a result, Usenet was great for short-term discussion and debate, but not much more. Moreover, although Usenet could be used to distribute files (such as pictures or software), it was awkward and slow to do so.
Perhaps more important, the software required to use Usenet, and the social conventions used to communicate with others were not easy to understand. As a result, it took a long time for newcomers to feel comfortable. Thus, as large as Usenet became, it satisfied only a small fraction of the need to connect.
By 1991, the ARPANET had been closed down, local area networks were abundant, and about 610,000 computers were connected to the Internet. In the spring of that year, a brand new information delivery system was introduced. It was designed to fit into a world in which the PC was ubiquitous and computing resources were distributed.
The new system, called the Gopher, was developed within the Department of Computer and Information Systems at the University of Minnesota. In April 1991, the first Gopher system was released by a team of programmers: Bob Alberti, Farhad Anklesaria, Paul Lindler, Mark McCahill and Daniel Torrey. Their goal was to provide a simple and inexpensive way for the various university departments to make information available to the campus at large.
The Gopher design used a hierarchical, file-like arrangement, with an interface that was instantly familiar to users. A gopher server would display a simple menu, in which each item represented either data (usually a text file), or a link to another menu. The syntax was simple, the overhead was small, and there was tremendous potential for expansion. As an example, Figure 1 below shows the most famous menu of the time: the main menu maintained by the Gopher programmers in Minnesota.
Internet Gopher Information Client v1.03
To access the system, you used a gopher client to connect to a server, which would display a menu. Every item on a menu was either some type of data or a link to another menu. You would select something, and your gopher client would fetch it from the appropriate server. If the item was a text file, your client would display it. If the item was another menu, the client would request it, show it to you, and then wait for you to make another selection.
The key to the Gopher system was that, from the beginning, it was designed to be distributed, with no central control. To join the system, all you had to do was set up your own server — usually referred to as a "gopher" — a process that was quick and inexpensive. As a user, you would access a gopher by using a gopher client program. Connecting multiple gophers was easy, because any link on any menu could point to a different server anywhere on the Internet.
For example, take a look at link #8 in the figure above: "Other Gopher and Information Servers". Starting with this link, you could navigate through a series of menus, and find your way to just about any gopher in the world. As you visited one gopher after another, you would be connected automatically to one computer after another, wherever it happened to be on the Net.
By the end of 1992, there were a great many gopher servers all over the Internet. This meant that users could — simply by following links — access information created and published by people all over the world. And yet, there was no central authority or clearinghouse. Every organization had control over its own information at a local level, and that information could be updated at any time. In this way, anyone with access to a gopher server, could publish information that would be instantly accessible to anyone with an Internet connection and a client program.
Is this starting to sound familiar? Have I not just described the basic concepts behind the Web? And yet, the Gopher system was conceived, created, distributed, and used widely several years before most people had even heard of the Web or hypertext.
Indeed, by the early 1990s, the Gopher system was enormously popular in many different countries. Thousands of gophers were established, and just about everyone believed that this was the system that would become the general Internet interface for everyone.
However, it was not to be, and therein lies a fascinating story.
The demise of the Gopher system as contender for "Most Popular Internet Service in History" was striking for three reasons. First, no one saw it coming. Second, it happened quickly. Third, it led directly to the Web, which did become the most popular Internet service in history.
In hindsight, it is possible to see why such a change was inevitable. The Gopher system was an ambitious client/server system that offered a simple, consistent interface to the vast resources of the Internet. Gophers were good at presenting well-defined chunks of information in a straightforward manner. This is why the system was an enormous hit within companies, universities, and government offices. Such organizations thrive on well-defined chunks of information and, philosophically speaking, they have a great need to disseminate it.
Individuals, unfortunately, do not communicate with menus. Nor are they satisfied when they are forced to access information using menus. (For example, how do you feel when you call a company, and you are forced to work your way through a phone menu system?) Put simply: the Gopher system required that all information be arranged into menus, and people just don't want to think that way.
What was needed was a new Internet system, one that could handle all the information of a gopher, while allowing individual people to create and share in a comfortable manner. The Internet needed something that would let anyone become a creator, while maintaining the order and the nurturing environment of a well-defined system. Such a something would encourage millions of new users to connect, induce thousands and thousands of organizations to develop a presence on the Net. It would make the Internet a permanent part of the popular culture, changing forever our view of mankind and our role in society.
Such a something was the Web: the fourth great attempt by our species to connect ourselves over the Net.
The concept that we now call hypertext is not at all new. It has been invented and reinvented numerous times. For example, in 1945, an engineer named Vannevar Bush discussed an information machine he called a "Memex". He observed that the amount of information available to our society was increasing rapidly and, in order to cope with such vast resources, readers would need to be able to follow various "trails" as they worked their way through the data. His Memex machine would help people do so.
Three and a half decades later, in 1981, an eccentric, iconoclastic writer named Ted Nelson self-published a book called Literary Machines, in which he coined the word "hypertext". He described a system named Xanadu, that would allow people to create and use hypertext.
However, the first popular hypertext computer application was not actually released until 1987. It was called HyperCard and was written for the Macintosh. HyperCard made it possible for Mac users to create and share "stacks" of information. Within each stack, there could be links from one data item to another. If you were using computers in 1987, you will remember how popular HyperCard became within a particular niche of the computing world.
As tempting as it might be to draw a line though Bush, Nelson and HyperCard, none of these inventions or ideas had anything to do with the Web (or even with one another). The history of science teaches us that, when the time is right, important new ideas will usually be developed separately by more than one person. For example, calculus was developed independently by both Isaac Newton (in England, 1665-1666) and by the Gottfried Liebnitz (in France, 1673-1676). Similarly, in the early 1980s, Tim Berners-Lee, a scientist at CERN in Switzerland, was developing his own idea of linked data, independent of the ideas of Bush and Nelson, and the principles of HyperCard.
Berners-Lee was a member of the high-energy physics community: the guys who spend enormous amounts of money building powerful machines to smash together small particles. CERN was a large laboratory in Geneva, Switzerland, devoted to the study of particle physics. (The word "CERN" is an acronym for the original name of the organization, Conseil Européen pour la Recherche Nucléaire.)
The physicists at CERN were awash in information, and Berners-Lee wanted to invent a method to make it all manageable. In March of 1989, he began to collaborate with another scientist, Robert Cailliau, on a project that would provide access over a computer network to what we now call hypertext.
Berners-Lee described what he was creating as a "web" consisting of a "network of links". The name stuck and, in May of 1991, the World Wide Web was released for use at CERN; in August 1991, the World Wide Web was announced on Usenet; and in January 1992, the first client program to access the World Wide Web was made available to the public via Anonymous FTP. This meant anyone with an Internet connection would be able to access the World Wide Web, simply by downloading and installing a client program.
To see what the Web looked like in 1992 take a
look at Figure 2, which shows the main Web
page at CERN. Compare what you see here to
the main Gopher menu in Figure 1. Notice
that, in both cases, the links (menu choices)
are identified by numbers. There were two
ways to follow a link. You would either type
a number, or you would move your cursor to the
link and press
If you were to select link #3 in Figure 2, you would jump to the page shown in Figure 3. This was a very well-known page, used as a starting menu by people around the world. From this page, it was possible to follow links #10 or #11 to search for information or to find other Web servers.
Berners-Lee and Cailliau did not invent hypertext, nor did they invent what we now call the Web. The idea of hypertext had been floating around for some time. For example, in November 1987, there was a conference called "Hypertext 1987" in Chapel Hill, North Carolina, at which many computer scientists presented papers. In July 1988, the ACM (Association for Computing Machinery) published eight of the papers in their journal Communications of the ACM. At the same time, they distributed a 5.25 inch floppy disk, called "Hypertext on Hypertext", containing an actual hypertext version of the papers. The disk used a system called HyperTIES, developed at the Human-Computer Interaction Lab at the University of Maryland. HyperTIES itself was based on work dating back to 1982.
The truth is, by the time Berners-Lee and Cailliau began their work, there were many people around the world working with hypertext systems. (Remember the idea we discussed earlier: most important new scientific concepts are developed independently by more than one person at a time.) So what did Berners-Lee and Cailliau do that was so noteworthy?
Three things: First, they were in the right place at the right time. Second, they developed a non-proprietary hypertext system that could be used by anyone. Finally, as the World Wide Web grew in popularity (which took some time), Berners-Lee and Cailliau participated in its development and insisted that it remain free and open. In this way, they were able to attract other people to work on the project, while ensuring that the fruit of their efforts would be distributed freely. As Berners-Lee observed in an interview in 2004, "Goodness knows, there were plenty of hypertext systems before [ours] that didn't interoperate."
At first the World Wide Web was a curiosity. By the end of 1992, for example, there were only 26 Web servers in the world. The basic idea of hypertext — data containing links to other data — had already been explored, and was widely accessible using the Gopher system. Moreover, the only Web clients in common use were text-based Unix programs, and Unix was difficult to use.
This all changed in February, 1993, when Marc Andreesen, a student at the University of Illinois, released Mosaic, the first World Wide Web client to take advantage of a GUI (graphical user interface): windows, scroll bars, a mouse, and so on. The first version of Mosaic ran under X Window, the GUI commonly used with Unix. This restricted its use to people who were fortunate enough to have access to a Unix workstation running X Window. Still, it was enough to show the world what the World Wide Web could look like with the right interface, and it changed people's thinking.
In the fall of 1993, two new versions of Mosaic were released, one for PCs running Windows, the other for the Macintosh. At once, everything came together:
• An information system that supported words, ideas, and pictures.
• Links that allowed you to jump from one item to another.
• A user interface that was familiar and easy to use.
• Client programs for PCs and Macs, the computers used by most people.
Now, anyone with Internet access could use the World Wide Web simply by selecting items with a mouse. Within a short time, use of the system began to grow and, by the end of the 1990s, the Web (as it came to be known) was dominant and the Gopher system was virtually gone.
What was so good about the World Wide Web? In the early 1990s it was far less sophisticated and not nearly as popular as the Gopher system. So why did the Web grow so quickly, and why did the Gopher languish?
The answer lies in the fact that the most important aspect of the Web was not the links that connect one data item to another. Rather, it was the text between the links. To see why this is so, compare the Gopher page in Figure 1 to the Web pages in Figures 2 and 3.
Gopher information was arranged in menus that were easy to understand and well-ordered. The problem is that people do not think in terms of menus. Web information consisted of words, sentences and (later) images, punctuated by links to other information. Although this is not exactly how people think, it is a lot closer.
When you used the World Wide Web, you looked at pages that contained text to read and (possibly) pictures to look at. When you used a gopher, most pages contained choices, not content. To actually get to something interesting — to read or look at — you had to navigate your way through a series of menus.
Thus, when you used the Gopher system, you had to separate the acquisition of ideas into two processes: selecting (from menus) and reading (text). The Web allowed you to blend these two processes into one continuous activity, an activity we now call "browsing". In hindsight, it is easy to see that human beings are more comfortable browsing than selecting and reading.
The Web was able to bring information to the surface, where it would be instantly accessible.
This observation explains why, ultimately, the Web came to dominate over the Gopher system, Usenet, and mailing lists. The basic design of World Wide Web — mixing content with links — was much richer than anyone realized. Unlike the Gopher system, the Web was able to bring information to the surface, where it would be instantly accessible.
By the mid-1990s, the efforts of countless people around the world had created a true wonder. What had once been the World Wide Web — a relatively simple client/server system for accessing hypertext from Web servers — evolved into the Web, an enormously complex information utility.
By 1995, it was clear that humanity had created its fourth great attempt to link people together over the Internet, and this time it was wildly successful, beyond anybody's expectations. By the end of the 1990s, the Web had fueled the growth of the Internet from an esoteric computer network to the grandest human creation of all time.
And yet— what is the real history of the Web?
Without mailing lists, we would not have had Usenet; without Usenet, we would not have had the Gopher system. And without all three, we would not have the Internet or the Web as we know it today.
If you want to explore the ideas we discussed in this essay, here are some resources you will find interesting:
Size of the ARPANET in 1979:
Number of computers (hosts) on the Internet
History of email:
Master list of Usenet Discussion Groups
Internet hosts: historical data from 1969:
History of the early Internet
Guide to gophers and other early Internet resources:
Early gopher main menu:
There are still gopher servers on the Internet. If you want to see what it's like to use a gopher server, here is an excellent place to start. If you use a browser that supports the gopher protocol use:
If your browser cannot connect using the gopher protocol, use the following address instead:
The "Hypertext 1987" conference:
The development of HyperTIES:
Tim Berners-Lee's proposal (March 1989) for the system that was to turn into the World Wide Web:
The address of the very first Web page was:
A copy of the first Web page viewed by most people in the early days of the World Wide Web:
Berners-Lee awarded the Millennium Technology Prize (June 2004):
© All contents Copyright 2021, Harley Hahn