The Machine in the Office
Tim Berners-Lee kept a NeXT computer in his CERN office with a warning written across it in red ink: “This machine is a server. DO NOT POWER IT DOWN!!” He wrote code directly on the computer, and the machine ran the first web server. Its address was info.cern.ch, a small point of entry that depended on one office machine remaining switched on.
Visitors who reached that address found the first website in the world, devoted to the World Wide Web project itself. The first Web page sat at http://info.cern.ch/hypertext/WWW/TheProject.html and explained the system through which it could be read.
A server in one office held information for other people, and every reader needed that machine to be reachable. CERN had built the Web to support automated information-sharing among scientists in universities and institutes around the world, so a powered-down computer could interrupt the point of the whole project.
Three years later, CERN would make the software available on a royalty-free basis, and the Web would spread far beyond that office. Why did Berners-Lee give the Web away after building it at CERN?
A Notebook of Connections
Berners-Lee arrived at CERN in 1980 for a fellowship, after studying physics at Oxford. During his work at the particle-physics laboratory near Geneva, he built a notebook program called “Enquire-Within-Upon-Everything”. It allowed links between arbitrary nodes, giving him a way to join pieces of information that had begun life in different places.
A person, a document and a piece of work could each become a node, then lead the reader onward through a connection.
In 1945, Vannevar Bush described a photo-electrical-mechanical device called Memex, designed for memory extension and for making and following links among documents on microfiche. Doug Engelbart’s oNLine System later brought hypertext browsing, editing, email and the mouse into a working system during the 1960s.
Ted Nelson gave this family of ideas its durable name in 1965, when he coined “Hypertext” in a paper presented at the Association for Computing Machinery’s twentieth national conference in New York. Andy van Dam and colleagues built the Hypertext Editing System and FRESS two years later.
Berners-Lee returned to CERN for a full-time job in 1984. The laboratory drew scientists and projects together, and the links in Enquire offered a practical way to keep track of relationships among people, documents and work. Five years later, he turned that practical idea into a proposal for the institution itself.
Vague, but Exciting
In March 1989, Berners-Lee circulated “Information Management: A Proposal” at CERN. It imagined a distributed information system with links, a simple client-server architecture and a six-month time frame. A client could ask for a document, a server could supply it, and a link could take the reader onward to another document held elsewhere.
Researchers at universities and institutes needed to share information automatically, and Berners-Lee described a system in which each participant could retrieve material from a connected source. The structure drew directly on the linked nodes he had explored in Enquire.
The people whom CERN served were already scattered across borders. CERN includes more than 17,000 scientists from over 100 countries, while the Web was conceived to meet demand for automated sharing among scientists in universities and institutes around the world. A distributed system gave such users a way to keep information with its source while making it retrievable elsewhere.
Mike Sendall, Berners-Lee’s boss, marked the proposal “Vague, but exciting.” CERN needed a way for scientists in many universities and institutes to share information automatically, and the proposal offered common machinery for doing it.
Berners-Lee resubmitted a slightly edited version in May 1990. Sendall then approved the purchase of a NeXT cube in September and allowed Berners-Lee to proceed with a global hypertext system.
Steve Jobs founded NeXT after leaving Apple, and the company sold only a few tens of thousands of devices. Berners-Lee used the NeXT’s development environment to build the early application in Objective C with Interface Builder, making the machine both his workshop and the home of a service for other people.
Writing the Web
Berners-Lee started work in October 1990 on a graphical hypertext browser and editor using NeXTStep. He called it WorldWideWeb, giving the same name to the program and the larger project. He had also considered “The Information Mesh” and “The Mine of Information,” names that described a store of material more than a place where readers could move freely among connected pages.
The program joined two activities in the same tool. A user could read hypertext and write it, which fit Berners-Lee’s view that a network of documents would grow through participation. He later put the point sharply: “If you think surfing hypertext is cool, that’s because you haven’t tried writing it.”
WorldWideWeb let people read and write hypertext. Berners-Lee wrote the first Web server, httpd, which ran in the background and connected to the browser, and he wrote the first version of HTML, the language used to structure Web documents. Around 1990, the browser, the CERN httpd server and the libwww C library developed in tandem.
Robert Cailliau joined the project in October as co-author of a new version of the proposal. In November, he and Berners-Lee submitted a formal management proposal called “WorldWideWeb: Proposal for a HyperText Project.” Their document gave CERN a description of software that was already taking shape on the NeXT.
The project now had a name, a formal document and a system that people could demonstrate.
The first server was initially called nxoc01.cern.ch, then became info.cern.ch. By Christmas 1990, Berners-Lee had defined the Web’s basic concepts, HTML, HTTP and the URL, and had written the first browser-editor and server software. The computer in his office had become both a working server and the host for the project’s own first site.
The Browser for Everyone at CERN
Nicola Pellow joined the project in November 1990 during her student work placement at CERN. Berners-Lee’s first browser-editor ran on the NeXT, while Pellow wrote a line-mode browser that could work with CERN’s central computers. Her task gave colleagues a route into the Web through the computers already available to them.
Pellow’s work addressed the immediate gap between the machine where the Web was being made and the machines on which CERN colleagues worked. The line-mode browser carried the same system to users whose computers could connect through text.
CERN later described the line-mode program as the first readily accessible browser for the Web.
By Christmas, the line-mode browser and WorldWideWeb browser-editor were demonstrable. In March 1991, the line-mode browser became available to users of CERN’s central computers. The program later worked with most Unix and Linux systems, extending the Web’s reach through a browser built for widely available machines.
The Web had started with a server and an editor on Berners-Lee’s NeXT, yet its purpose depended on people who could retrieve material from their own computers. The line-mode browser brought that purpose closer by making the system available to CERN colleagues.
Berners-Lee released his WWW software in 1991, including the line-mode browser, server software and a library for developers. The release gave people the parts required to read Web pages, run a server and build on the system.
Other Machines Join
Berners-Lee announced the software on Internet newsgroups in August 1991, opening the project to people beyond CERN. The announcement offered a way into a system whose value increased as more people could run servers and publish linked material. Each new installation gave readers another destination and gave its operators a way to join the shared network.
Paul Kunz and Louise Addis carried that possibility to California. On 12 December 1991, they brought the first Web server outside Europe online at the Stanford Linear Accelerator Center. The server at SLAC meant that Web information could live in another institution while remaining accessible through the same links and protocols.
Kunz and Addis brought the server online through their efforts, so a reader could use the same Web system to reach material at a California laboratory and at CERN. That arrangement fulfilled the client-server model in the proposal: information stayed with separate institutions while links made it reachable.
CERN’s original server remained at info.cern.ch, while SLAC could maintain material on its own server in California. Readers could move between the institutions through the same Web system, because the project had been built for information held in many places.
Each server could hold its own documents, each client could ask for them, and links could connect material across the network. A single central store would have preserved control in one place; Berners-Lee’s proposal had instead imagined information distributed across many places.
By late 1993, there were more than 500 known Web servers, and the Web accounted for 1 percent of internet traffic. The system had grown from one labeled NeXT computer to a network of machines maintained by separate organizations. CERN then faced the decision that would determine whether those organizations could keep joining on equal terms.
The Release
On 30 April 1993, CERN made the source code of WorldWideWeb available on a royalty-free basis. CERN also put the World Wide Web software in the public domain and later made a release available with an open licence. Anyone could use the technology without paying fees to CERN.
CERN had made the technology freely usable by anyone, with no fees payable to the laboratory. The choice put every prospective operator in the same position as SLAC: able to install a server and offer material through the shared system.
The release followed the project’s original purpose. The Web had been conceived to meet scientists’ need for automated information-sharing across universities and institutes, and its basic parts depended on shared use: servers needed to understand requests from browsers, documents needed common markup, and links needed addresses that readers could follow. Fees and exclusive control would have placed barriers inside a system designed to connect separate places.
Adoption accelerated. By the end of 1994, the Web had 10,000 servers, including 2,000 commercial servers, and 10 million users. The same year, Robert Cailliau organized the first International World-Wide Web Conference at CERN, drawing 380 users and developers. A second conference in the United States that October drew 1,300 people.
Cailliau had helped turn Berners-Lee’s proposal into a formal project at CERN. The conference showed the people now working with its results: users and developers gathered at the laboratory where the first server had run. By the second conference, the gathering had moved across the Atlantic and grown substantially larger.
Berners-Lee left CERN for MIT in 1994 and founded the World Wide Web Consortium with CERN, DARPA and the European Commission. Companies had begun investing increasing resources in the Web and urged the creation of the consortium. The work of coordinating a shared system had become larger than any one laboratory, even as its essential decision had already been made: people could use the underlying software freely.
The Red-Ink Instruction
The release answers the question in the terms Berners-Lee had set from the start. He built the Web for information-sharing among dispersed scientists, and CERN made its software free for others to use because the system depended on many people and institutions being able to connect, publish and follow links. The first site at info.cern.ch had explained that project from inside the project itself.
By the end of 1994, a reader could reach Web material through thousands of servers instead of relying on Berners-Lee’s office machine. Yet every server still performed the basic job of the NeXT computer: it kept information available for someone elsewhere to request. The Web expanded by repeating that modest arrangement across universities, laboratories, companies and homes.
The handwritten label on Berners-Lee’s NeXT now reads like a compact instruction for the network that followed. The machine held the first server, hosted the first website and carried the code from which a wider system grew. Its red letters asked colleagues to keep one connection alive: “This machine is a server. DO NOT POWER IT DOWN!!”
Comments (0)