IN DECEMBER 1991 PAUL KUNZ, a particle physicist at the Stanford Linear Accelerator Center in Menlo Park, CA, programmed the very first Web server in North America. I learned about it the following January, upon returning to SLAC from a year away in Washington, DC. “Have you seen the latest?” an enthusiastic colleague asked. “It’s called the World Wide Web!”
At the time, 30 years ago, I didn’t grasp the tremendous historical significance of this event. Hardly anybody did.
This was the pivot point at which the Web truly became a worldwide system. Until that December, it was a European phenomenon focused at the European Center for Particle Physics (CERN) near Geneva, developed there by physicist-turned-programmer Tim Berners-Lee. In 1991 physicists who came to and worked at CERN from throughout the continent had begun to use the Web to share documents and data. An early “killer application” (or killer app) was the CERN phonebook, which could now be accessed via the Web from different computer platforms at the lab or externally via the Internet.
The SLAC server, set up on its mainframe computer, allowed physicists everywhere to access its valuable SPIRES database of preprints — as-yet-unpublished papers that had been submitted to scientific journals. Before that, it had been difficult to access this collection from beyond SLAC. Afterwards, accessing it became easy. Web traffic tripled in the two months after Kunz established the SLAC server, making SPIRES its new killer app. “A world of information is now available online from any computer platform,” boasted Berners-Lee in CERN’s computer newsletter that December, “Information sources at CERN and across the world span subjects from poetry to biochemistry and supercomputing.”
Physicists then were largely unaware of its likely impacts, but a major rupture in world history occurred late that month. Meeting in Moscow on 26 December 1991, a day after Mikhail Gorbachev had resigned as president, the Supreme Soviet voted to dissolve the Union of Soviet Socialist Republics. The bipolar world order of the Cold War, which had been disintegrating since the fall of the Berlin Wall in November 1989, officially ended, to be replaced — at least during the next decade, before 11 September 2001 — by one in which democratic Westernized nations predominated, knit together increasingly via the Internet and Web.
“There can be no serious doubt that, in the late 1980s and early 1990s, an era in world history ended and a new one began.”
— Eric Hobsbawm
As British historian Eric Hobsbawm observed in his epochal history of the twentieth century, The Age of Extremes, “There can be no serious doubt that, in the late 1980s and early 1990s, an era in world history ended and a new one began.” Physicists and programmers — especially at CERN and SLAC — were to play important roles in shaping the technological foundations of the emerging globalized world order.
Part One: Origins of the Web
The origins of the Web can be traced to the ARPANET, which emerged in the 1960s during the depths of the Cold War. This was a project of the US Defense Department’s Advanced Research Projects Agency to build a decentralized, nationwide computer-communication system that could survive major disruptions (e.g., in a nuclear war) and continue functioning. During the 1980s, early computer networks such as BITNET, DECnet, ESNET and NSFNET emerged in the United States; by adopting ARPANET’s Transmission Control Protocol and Internet Protocol (TCP/IP), they permitted computers to communicate across one another and led to the “network of networks,” or Internet.
CERN had used TCP/IP for its internal networks, but it initiated external connections to the Internet only in January 1989. It proved the ideal site for the Web’s germination. During the 1980s, hundreds of physicists were arriving there from all over Europe — and increasingly from the United States — to do research on its Large Electron-Positron (LEP) collider and detectors. They brought with them a variety of computers, which communicated with one another over diverse systems and networks set up by the various experimental collaborations. Most would return to their home institutions, but they still needed to communicate via their computers with colleagues at CERN and elsewhere.
Enter Tim Berners-Lee, a British physicist turned computer programmer, who had first worked at CERN in 1980 as a consultant on a particle-accelerator control system. He returned in mid-decade as a CERN Fellow, which allowed him the freedom to follow his personal instincts in creating software. Thus, two months after the Internet reached the lab, he began to advocate marrying it with the point-and-click techniques of hypertext in a March 1989 memorandum titled “Information Management: A Proposal. [1]”
“CERN is a model in miniature of the rest of the world in a few years’ time,” he wrote. The lab “meets now some of the problems the rest of the world will have to face soon.” By marrying hypertext and the Internet, he envisioned, users could point to the text of references on a computer screen and “skip to them with the click of a mouse.” [1]
The following year he teamed with Belgian software engineer Robert Cailliau, who had been working at CERN since the 1970s and then headed its Office of Computing Systems. Also interested in hypertext, he had much more experience obtaining CERN resources. Together they conceived an audacious name for their initiative and in October 1990 issued a document titled “World Wide Web: Proposal for a HyperText Project.”
Berners-Lee and Cailliau had been able to procure expensive NeXT computers — an innovative workstation marketed by the firm Steve Jobs founded after leaving Apple in 1985. It had a graphical user interface (GUI) and powerful programming software that helped Berner-Lee to develop the Web software in a few creative months, including core features we now take for granted, such as the HyperText Transfer Protocol (HTTP) and HyperText Markup Language (HTML). One computer would act as the “server,” providing useful digital information while other computers acted as “clients” or “browsers” seeking this information. By Christmas 1990, they had the system up and running, with the world’s first Web server (nxoc01.cern.ch) residing on Berners-Lee’s NeXT computer and Cailliau accessing it from his machine.
But given their $6,500 cost (equivalent to about $15,000 today), NeXT computers were extremely rare in high-energy physics, well beyond the financial resources of nearly all in the field. So the two men encouraged a young CERN intern, Nicola Pellow, to program what they called a “line-mode browser” — a simpler Web interface than the NeXT computer’s elaborate point-and-click GUI browser. It could be employed on many other computers used at CERN, so that “no matter what machine someone was on, he would have access to the Web,” recalled Berners-Lee. [2] And he succeeded in getting a Web gateway established to the CERN phonebook database, providing easy access to an information resource that nearly everyone at the lab used almost daily. “Mundane as it was,” wrote Berners-Lee a decade later, “this first presentation of the Web was, in a curious way, a killer application.”
“By 1990 CERN had become the largest Internet site in Europe,” recalled Ben Segal, who had served as the lab’s network czar during the 1980s. The laboratory “positively influenced the acceptance and spread of Internet techniques both in Europe and elsewhere.” [4] And the Web soon was the major reason for institutions to gain access to the Internet. In March 1991, for example, the Internet became accessible in Czechoslovakia, Hungary and Poland after the US National Science Foundation extended the NSFNET backbone into these Eastern Bloc countries. And the previous year, Russian physics labs had established a link to the Internet via direct phone lines to Finland (this link surreptitiously kept the outside world appraised of events in Moscow during the August 1991 Soviet coup). European physicists involved in CERN research were thus among the earliest to take advantage of Internet access, many of them via the Web. In networking and software, CERN led Europe onto the emerging “information superhighway.”
That was a popular name for the Internet championed by Vice Presidential candidate Al Gore during the 1992 US national elections. As a Tennessee Congressman and Senator, he had foreseen the vast economic, social, and cultural impacts of the new network technology. And he helped enable its further evolution by sponsoring the High-Performance Computing Act of 1991, which the US Congress passed that autumn and President George H. W. Bush signed into law on 6 December 1991. It was undoubtedly a pivotal month in world history.
Three months earlier, Paul Kunz had passed through CERN on his way home to SLAC from Sweden and met there with Berners-Lee, who showed him the Web. Kunz quickly grasped its potential after seeing how it could be used to communicate over the Internet with his NeXT workstation back at SLAC. “I was really excited,” he recalled. “I told Tim I was going to put SLAC’s SPIRES database on the Web as soon as I got home.” [3]
Just after returning, Kunz visited SLAC librarian Louise Addis. “I’ve just been at CERN and found this wonderful thing a guy named Tim Berners-Lee is developing,” he told her. “It’s called the World Wide Web, and it’s just the ticket for what you guys need for your database.” But it took more than two months to set it up, in part because Kunz delegated the work to SLAC programmer Terry Hung and because the server was to be established on the SLAC IBM-VM mainframe computer. So Addis and Berners-Lee pressed Kunz back into service to finish the job.
On 12 December 1991, the first Web server outside of Europe (slacvm.slac.stanford.edu) went online, making SPIRES available throughout the world via the Internet. Less than 24 hours later, Berners-Lee announced its existence it on the WWW interest-group bulletin board: “There is an experimental W3 server for the SPIRES High Energy Physics preprint database, thanks to Terry Hung, Paul Kunz and Louise Addis at SLAC.” [3]
Exactly two weeks later, the Soviet Union collapsed, and the Cold War officially ended. Few physicists recognized the global significance of this epochal event, however, continuing their research as if nothing important had happened. The Superconducting Super Collider (SSC) Project in Texas, for example, was then preparing to ask Congress to approve a $2.35 billion increase in its construction cost, to $8.25 billion.
By the summer of 1992, three more high-energy physics labs had set up Web sites: the Fermi National Accelerator Laboratory (Fermilab) west of Chicago, the Deutsches Electronen Synchrotron (DESY) lab in Hamburg; and the Dutch National Institute for Subatomic Physics in Amsterdam. And the MIT Laboratory for Computer Science, where Berners-Lee visited that summer, established a server in June. Others soon followed as he and Cailliau hit the lecture circuit to promote their networking software.
A signal event was the 1992 Computing in High-Energy Physics Conference held in Annecy, France, that September. There they gave Web demonstrations and a prominent lecture titled “World Wide Web,” in which they showed a map of the rapidly growing number of Web servers in Europe and North America, plus others planned for Asia and Australia — then about 20 in all. The presentation was the principal highlight of the gathering for Terry Schalk, a particle physicist at the University of California, Santa Cruz, who in summarizing the meeting, remarked that “if there is one thing everyone should carry away with them from the conference, it is the World Wide Web.” [5]
By November 1992 there were at least 42 Web servers worldwide, most of them at high-energy and nuclear physics laboratories in Europe and the United States. What was needed then for the Web to gain much wider acceptance were GUI browsers that functioned on the personal computers and workstations then commonly in use. Cailliau and Pellow had been programming such an Apple Macintosh browser on and off for almost a year, and a group of students working with them was developing an IBM-PC browser. But these fledgling efforts suffered from the lack of high-level management support.
Thus it was CERN outsiders who ended up developing the first useful GUI browsers, at Cornell, UC Berkeley and SLAC. In the summer of 1992, physicist Tony Johnson — a member of an informal group calling themselves the “SLAC WWW Wizards” — was developing such a browser for computers with Unix operating systems, which he dubbed Midas (or MidasWWW). [5] Like the GUI browser that Berners-Lee had programmed on his NeXT workstation, it could display images in separate windows. Midas was becoming popular at SLAC, but Johnson was initially reluctant to make it widely available, given the additional work that would be required to maintain it. He finally released the browser on the FREEHEP server in mid-November, announcing it to the growing Web community on the WWW-talk newsgroup.
Among the first to download this software and try it out was Marc Andreessen, a young computer-science undergraduate at the University of Illinois. He had become enamored of the Web that month while working at the National Center for Supercomputer Applications (NCSA) on campus, developing computer-visualization software. He emailed Johnson about it the next day. “Midas WWW is superb!” he exclaimed. “Fantastic! Stunning! Impressive as hell!” [3]
Andreessen suggested they collaborate on an improved version incorporating graphic files, animations, and other advanced features. Intrigued at first, Johnson turned down the offer after considering how such an intense programming activity would conflict with his physicist day job. So Andreessen teamed instead with NCSA staff member Eric Bina to develop an X-Windows browser for Unix-based computers. After a feverish two months of programming activity that included many pizza-fueled all-nighters, they released what they dubbed NCSA X Mosaic 0.5 on 23 January 1993 (three days after the Clinton-Gore inauguration). “Brilliant!” Berners-Lee exclaimed on the WWW-talk newsgroup. “Having the thing self-contained in one file makes life a lot easier.”
Over the next two months, Andreessen and Bina released several more versions of X Mosaic in rapid succession. By early March, their browser could include embedded graphics on computer screens for the first time — rather than having to open separate windows to display them. This was a pivotal salient into commercial territory. Whereas early Web pages had been dominated by text (with underlined text for hypertext links), they could now resemble glossy magazine pages. The advertising industry began to take notice.
That April NCSA officially released X Mosaic version 1.0. But Unix-based computers were largely the domain of the academic community or government and industrial laboratories. To commercialize the Web, companies and ordinary users needed browsers that functioned on Macintosh and IBM-PC computers. Other NCSA programmers were developing such browsers — which could also support embedded graphics. In the autumn of 1993, NCSA released Mosaic browsers for Mac and IBM-PC computers to great fanfare. These easy-to-install browsers, with superior graphics capabilities, were soon being downloaded by the thousands every month. (Now I could finally access the Web from my PowerMac computer, not just on the SLAC mainframe.)
The New York Times was among the first to pick up on the news among the mainstream media. “A new software program available free to companies and individuals is helping even novice computer users find their way around the Internet,” announced technology reporter John Markoff on the front page of the 9 December 1993 Business Day section. “Since its introduction earlier this year, the program, called Mosaic has grown so popular that it is causing data traffic jams on the Internet.”
Together the Mosaic browsers and Web provided ready on-ramps to the information superhighway for the rapidly expanding legions of personal-computer users. With a burgeoning audience now available, the number of Web sites literally exploded, to more than 500. By the end of December 1993, even the White House had one.
“By 1994 the centre of gravity of the World Wide Web had crossed the Atlantic to a place where the entrepreneurial heart beats stronger,” observed Cailliau and James Gillies in their 2000 book, How the Web Was Born. [3] To be more exact, it had shifted to Silicon Valley — into which SLAC physicists had introduced the Web two years earlier.
Although few high-energy physicists anticipated the Web’s commercial potential, Marc Andreessen surely did. In early 1994, he headed west to Silicon Valley, where he joined venture capitalist Jim Clark in establishing Netscape Communications Corporation and recruiting fellow NCSA programmers, including Eric Bina. By year’s end, this team had developed the Netscape Navigator browser, which eclipsed Mosaic within a few months, becoming the dominant Web browser. After the firm went public in August 1995, its stock value quickly tripled to $58, giving it a market value of nearly $3 billion and making Andreesen a millionaire many times over.
CERN workers had pioneered an early keyword-search function and the WWW Virtual Library of interesting Web sites, but these were soon swamped by commercial operations established by Stanford University graduate students who became early adopters of the new networking software that had touched down at SLAC a few years earlier. In 1994 electrical-engineering graduate students David Filo and Jerry Yang began compiling an online list of their favorite servers, which they dubbed Yahoo! after hundreds were accessing it regularly. Soon it was experiencing millions of hits a day, so they found a local venture-capital firm to provide money and management know-how and to take the company public — which happened in April 1996, to the tune of $34 million.
Several early attempts to establish Web search engines, each with its own strengths and weaknesses, had proved frustrating. But in 1995 to 1996, two Stanford computer-science graduate students, Larry Page and Sergey Brin, hit upon a novel idea: to rank a Web site according to the other sites linked to it. Their “PageRank” algorithm formed the basis of the new Google search engine, originally released on Stanford University’s Web site in August 1996. Incorporated in 1998, the firm Google, Inc. (now Alphabet, Inc.) has since come to dominate searching and other aspects of the Web. The corporation finally went public in August 2004, at a market value of $23 billion.
A common element in this explosion of commercial Web activity was the proximity of the Valley’s venture-capital industry headquartered just across Sand Hill Road from SLAC in the Sharon Heights district of Menlo Park — for example, Kleiner-Perkins, Menlo Ventures and Sequoia Capital. Technology financiers at these firms began viewing computer software as a high-growth, high-tech industry after the 1980s successes of Apple and Microsoft. With innovative Web-based software firms bursting out literally at their feet, they were quick to fund them with much-needed capital and the management expertise required to promote rapid growth and to take them public.
Nothing like this venture-capital industry existed near CERN, despite several major international banks headquartered in Geneva and other Swiss cities. And to my knowledge, no SLAC WWW Wizard ever crossed Sand Hill Road to seek venture financing. Just steps away, this Sharon Heights financial community was a distinctly different culture from the atomic and high-energy physics research efforts consuming SLAC scientists, engineers and programmers daily. Thus high-energy physicists largely missed the vast commercial promise of the Web (and in certain prominent cases resisted it).
By 1994 Berners-Lee could see the handwriting on the wall, especially after the first International Conference on the World Wide Web held at CERN that May, which attracted nearly 400 attendees and was dubbed “the Woodstock of the Web.” With commercialization came the possibility of the Web’s fragmentation, as powerful business interests tried to pull it in specialized directions more amenable to individual profit motives. He too crossed the Atlantic that year, in September, to assume a position at MIT’s Laboratory for Computer Science as director of the World Wide Web Consortium, which involved both academic Web users and commercial purveyors of Web software and services. [3] CERN had initially agreed to serve as the European headquarters of the consortium, but it stepped aside in favor of the French computer-science organization INRIA after deciding that December to proceed with the costly Large Hadron Collider project. After that, the principal role of high-energy physicists in fostering the Web’s further evolution became that of pioneering its use in the coordination and management of global-scale activities and projects. (see “Going Global 2“).
Top Photo: Paul Kunz (left) in his SLAC Office with Tim Berners-Lee in 2000, with Paul’s NeXT computer between them. (Courtesy of SLAC)
A condensed version of this article, titled “Going Global: What the Web Has Wrought,” has been published in the January 2022 issue of Physics World.
References
1. Tim Berners-Lee, “Information Management: A Proposal,” March 1989.
2. Tim Berners-Lee with Mark Fischetti, Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web. (New York: Harper Collins, 1999).
3. James Gillies and Robert Cailliau, How the Web Was Born: The Story of the World Wide Web. (Oxford, UK: Oxford University Press, 2000).
4. Ben Segal, “A Short History of Internet Protocols at CERN.”
5. Bebo White, “The World Wide Web and High-Energy Physics,” Physics Today (November 1998).
Orcas Island physicist and writer Michael Riordan is author of the award-winning 1987 book The Hunting of the Quark and coauthor of The Solar Home Book (1977), The Shadows of Creation (1991), Crystal Fire (1997) and Tunnel Visions (2015). His articles and essays have appeared in the New York Times, Seattle Times, Scientific American, and many other publications. He serves as Editor of Orcas Currents.
I truly enjoyed reading your excellent history of the World Wide Web, Michael! The World Wide Web transformed the biological sciences during my career at the University of California at Berkeley.
Thanks, Janet. It turns out that UC Berkeley also had a role to play in the early history of the Web, but I did not mention it here. A student named Pei Wei programmed the ViolaWWW (or just “Viola”) browser in 1992 for Unix-based computers. Like Tony Johnson’s Midas, it influenced Marc Andreessen in the development of the NCSA Mosaic browser during the winter of 1992-1993.
And the rest, as they say, is history!
The WWW—Tim-Berners-Lee—CERN story reminds me of
the arXiv—Paul-Ginsparg—Los Alamos story involving another physicist who, by establishing the online preprint depository arXiv for physics (and mathematics) at Los Alamos National Laboratory, made physics research more democratic. Ginsparg was recently given the first Einstein Award from the Einstein Foundation in Berlin (see https://www.einsteinfoundation.de/en/award/) for his contribution. The physics/mathematics arXiv now lives at Cornell University.