By William Saletan '87


It's one o'clock in the morning, and Theodor Holm Nelson '59 is explaining in a blizzard of mind-bending metaphors how he conceived the origins of what is now the World Wide Web. Nelson, widely regarded as the founding prophet of interactive media, is speaking to me by phone from the other side of the world. It's midday at Keio University in Japan, where Nelson teaches, writes, and designs software, but he's exhausted. He works at night and sleeps during the day. "Mine is a parallel universe," he explains on his home page on the Web. "I share the physical universe with other people, but it seems I see it very differently. So my world is the same but different."

And how. Everything about Ted Nelson is upside down. While others study how things work, Nelson imagines how they could work differently. While others adapt to systems, Nelson adapts systems to people. While others plan forward from the status quo, Nelson plans backward from perfection. That's why he foresaw the age of cyberspace four decades ago and why he is so dissatisfied with it today.

The first inklings of interactive media came to Nelson in childhood. When he was 5 years old, he contemplated how florists sent flowers by wire. "What did they do to the flower," he wondered, "so that they could send it down the wire and rebuild it at the other end?" His parents' show business careers also inspired his imagination. His mother, Celeste Holm, starred in more than a dozen movies and won an Academy Award for Best Supporting Actress (for her role in Gentlemen's Agreement) when Nelson was 10. His father, Ralph Nelson, directed 11 films and won an Emmy for directing the television production of Requiem for a Heavyweight in 1956.

"When I was 12 years old, bang, here came this thing called television," Nelson recalls. "So the notion of one medium succeeding another was plain to me, as was the notion of men at work in front of cathode ray tubes. I sat behind my father [in the studio]. These bright, intelligent guys sitting at screens with absolutely captive intensity in the control rooms--that seemed to me the way it ought to be." When the Japanese film Rasho-mon came out in 1949, Nelson saw it three times. Its portrayals of a jointly witnessed story intrigued him. Film and literature, he realized, could be presented not merely in a preordained sequence but in "multiple parallel versions."

Nelson arrived at Swarthmore allergic to orthodoxy of all kinds. "I majored in extracurriculars," he says. He challenged the College's restrictions on fraternization between male and female students. "In the 1957 student affairs committees, sitting there with Dean Susan Cobbs, who hated my guts and tried to get me expelled, I was arguing for sexual freedom--before anybody else dared open their yaps about it," he recalls. Nelson was also an incorrigible prankster. One time, as Nelson remembers it, Dean W.C.H. Prentice gave a Collection speech exhorting students to turn in "troublemakers" in their midst. "The following Monday, in everyone's mailbox there was a typeset confession blank, saying that because the deans did not wish to maintain an elaborate police system, 'Your cooperation is expected.'" Below the words, "I have recently committed the following offenses," the form listed various categories and a line for the student's signature. "There was a firestorm of protest," Nelson laughs. "A couple of hundred people sent them in."

In the classroom, Nelson criticized theoretical constructs, arguing that they oversimplified reality. He didn't even like the paper they were printed on. Ideas were connected in multiple dimensions, he reasoned. So trying to represent them on paper, much less edit them, was intrinsically crude. "There was no way, with paper, to represent the changes and structures and connections," he explains. "You can represent the changes with arrows, but after a certain point, you have to retype it." This led Nelson to the idea of hypertext. "David Rose ['60] told me that I laid out the whole hypertext idea to him when we were undergraduates together."

Nelson's ideas crystallized a year after graduation, when he enrolled in a computer course at Harvard. He started with the idea of word processing. As Nelson saw it, word processing (which was yet to be invented) would allow the writer to revise each draft. Nelson wanted a program that would also link each section of the new draft to the same section of previous drafts--as far back as the original notes. By backtracking through these links, the writer could compare serial versions of each section and could recombine current and previous versions of the various sections into a new draft. The program remained unfinished, but Nelson's dream of linked literature grew. In 1965, he gave it a name: "hypertext."

Hypertext, as Nelson conceived it, would not be limited to fixed sequen-ces--as paper required--but would allow readers to move within a text, or between texts, in whatever sequence they fancied. Suppose that while reading a history of Swarthmore, you came across a reference to Crum Creek. You might choose, with a mouse click, to veer away from the College's history and explore a tangent string of literature about the Crum. On this path, you might come across a reference to Spiro Agnew's description of Swarthmore as "the Kremlin on the Crum." Whereupon you might choose to go off on a tangent about Agnew, and so on.

A hypertext system needed digital addresses for documents, so users could find them through links. But Nelson wanted much more. He wanted users to be able to create links in documents written by others. Nelson also wanted the system to track every change in each document and automatically to reorganize the links accordingly, which was immensely more difficult. Nelson was seeking nothing less than a universal, self-updating library. He called it Project Xanadu.


For more than a decade, Nelson's Swarthmore connections helped nurture the project. Because his ideas were radical, "The conventional computer establishment locked me out," he says. "At Swarthmore, there were people who could help." Nelson visited the campus frequently from 1970 to 1972, drawing on the advice of faculty friends and the assistance of two of Swarthmore's first computer science students. In 1976, the College invited Nelson back to teach courses in interactive software and hyper-media.

Nelson's courses brought him new disciples, and in the summer of 1979, he summoned his followers from around the country to Swarthmore to complete the Xanadu project. He was no longer teaching on campus, but he enjoyed the atmosphere and maintained an apartment nearby. His disciples rented a house and redesigned the entire system from scratch. "We were pushing the envelope of practicability on every side," says Nelson. "We were writing a program that was bigger than compilers could handle." Eventually, the programmers ran out of money, and Nelson and his group joined a software company in Texas. Xanadu was put on hold. The "Swarthmore summer," as Wired magazine referred to it years later, was "Xana-du's golden age."

But while Xanadu lay dormant, its seeds took root. In 1987, Apple introduced HyperCard, a program that al-lowed users to construct webs of links within their personal computers. The program's name and concept transparently derived from Nelson's work. "HyperCard apparently came out of the talk I gave at the hackers' conference in 1984," he says. "I was invited to fly to Apple in '86 or '87 for a chat with [Apple CEO] John Scully.... I think he wanted to take my measure and see whether I would object to the word 'HyperCard' being used for what they were about to release." When the program came out, Nelson recalls, "Everybody said, 'Oh, my god, Nelson, this is what you've been talking about!'"

Not really, says Nelson. The problem with HyperCard was that it was stuck inside a single machine. Nelson's dream required a network so that users could link to documents on other people's computers. The Internet solved this problem. In 1989, Tim Berners-Lee, a physicist and former software developer, began to build a system for reading and writing hypertext on the Internet. He called it the World Wide Web. Berners-Lee "dropped by my office in '92 when the [Web] was in alpha or beta," says Nelson. "He wanted to show it to me because he had just heard about my work." Nelson, it turned out, had been the source of many of the words and ideas that had filtered down through the software world and were now coming to life on the Web: links, a digital registry of document addresses (uniform resource locators, or URLs), a "hypertext transfer protocol" (HTTP), and a "hypertext markup language" (HTML).

Nelson also influenced the transformation of the Web into a mass medium. In 1989, he presented his Xanadu design to the National Center for Supercomputing Applications (NSCA) at the University of Illinois. Larry Smarr, NSCA's director, was sufficiently impressed that he expressed interest in joining the Xanadu project. Later, Smarr told Nelson he had begun a new project of his own. That turned out to be Mosaic, the first consumer-oriented Web browser, which was produced by Smarr's lieutenant, Marc Andreessen, and released in November 1993. Six months later, Andreessen left NCSA to found Netscape Communications, which turned the Web from a communications tool of scientists into a popular interactive medium.


You'd think Nelson would be happy to claim the Web as his brainchild, but he isn't. While others celebrate what it can do, he dwells on what it can't do. You can't write notes on somebody else's document or overlay them on the original. You also can't quote significant parts of somebody else's document without violating copyright law. And when you read one document that quotes another, you can't see the original context without finding your way to the quoted document.

Nelson has always thought to solve these problems with an idea he first named "transclusion," but now calls "hypersharing." Every document would be linked intimately to every other document that quotes it or refers to it, so that the reader could see them simultaneously. Suppose Nelson wanted to criticize the writings of Dean Cobbs. While reading Cobbs' work, you could see Nelson's comments alongside the original. And conversely, while reading Nelson's comments, you could see Cobbs' words, so Nelson wouldn't get away with taking them out of context.

Nelson goes back to the premise of his student days: Computers should improve on paper. "What they're giving us now is so much less than paper, it's pathetic," he complains. "One of the things you can't do on a computer screen is flip. With a book or stack of cards, you can flip through at great speed, and your eye sees things. Opening a file on a computer is like prying open a wooden crate. It's absurd." The more Nelson talks, the more you realize that what's nutty isn't his analogy, but the fact that you've never noticed the ways in which your computer is inferior to the stuff your parents wrote on. For example, he observes, "Paper is compatible." Furthermore, "if I buy a CD-ROM, I'm not allowed to write in the margins. What the hell is that about?"

Defenders of the Web, CD-ROMs, and other imperfect technologies say they've engineered real progress while Nelson has held out for idealistic "vaporware"--software that is promised but never delivered. Some think Xanadu expressed the radical political ambitions of the 1960s and 1970s and ultimately collapsed for the same reasons. Nelson always envisioned technology that would promote equal access to information and freedom of the press. Even his metaphors are populist and countercultural. In a recent essay on piracy and copyright law, he argued: "There is a hunger for the reuse of media. If we can find a legitimate way to feed this hunger, then perhaps the stealing will not be necessary." He calls his latest software "the sexual revolution brought to the spreadsheet. Spreadsheets require that a cell have an up connection, a down connection, a left connection, and a right connection. In my system, each cell's connections are its own business."

Much of the criticism of Nelson's ideas echoes criticisms of the political left. One reviewer has called Xanadu a "weird, semialtruistic/semifascistic vision" of a world in which "all information is irrevocably networked." Others have portrayed Nelson as naïve for supposedly postulating a "communal database administered, of course, from above." These skeptics prefer the Web and its free-market patchwork, in which you can revise or remove anything you've written without being monitored and without concern for others who may have linked their work to yours. It's better to let individuals work out these occasional breakdowns, goes this argument, than to hold out for a perfect system.

Nelson's rejoinder, in turn, echoes criticisms of the political right. If the Web's defenders think the cybermarketplace is truly free, they're deluding themselves, he suggests. "All information is already irrevocably linked," says Nelson, "but the connections are currently hidden. Some people want to keep things hidden; I don't."

The opposite case against Nelson's vision is that it's too anarchic. He derides "regularity chauvinists" and rejects traditional software concepts such as "files," "icons," and even "metaphors." Instead, he proposes "a totally explorable, customizable, user-reconfigurable, and shareable world." But in a such a world, wouldn't people get lost? Skeptics use Nelson's personal life to illustrate the point. He is notoriously disorganized, blessed and cursed with a brain whose rapidly branching trains of thought defy conventional notions of coherence. His chaotic books and manual filing systems are legendary. When I interviewed him, his unrelenting tangents and objections to the premises of my questions made a hash of the outline I had prepared. "My world is not organized around your outline," he later tells me.

Nelson concedes that many computer users prefer the comfort of dictated structure to the responsibility that comes with total freedom. "I'm not interested in that trade," he replies nonchalantly. "I'm not interested in pandering. I want to make people aware of the necessity of freedom." Structure is useful, he argues, but only if it's "permeable" and open to change by the user. "I believe in creating elegant environments that help and encourage people to create the structure they want. But that doesn't mean it has to be some particular restrictive structure that the dweebs created in the last 20 years." As to the notion that total freedom is scary, Nelson brushes it off with his favorite analogy: "It's like a blank piece of paper."

The conventional wisdom in the computer industry is that Xanadu failed, that more practical software makers succeeded, and that Nelson and his ideas are relics. Nelson returns this condescension by dismissing the Web as irretrievably mediocre. "I don't think you can put wings on a child's wagon," he scoffs. But beneath this posture, Nelson has quietly reconsidered his past and his future. "To me, binding the whole thing together into an indivisible structure was always the center. Being a monist made it hard to break it down into tactical goals. And that's a fundamental failing that's left me where I am. If I were a company man or somebody able to do small things in a small way, it would have been much more effective."

Accordingly, Nelson has changed approach. He has divided Xanadu into pieces that can be grafted onto the Web. "Now that we have a dispersed Web under dispersed ownership and management, we have to create systems that can be marketed on that Web," he explains. One piece is a "micropayment" system under which each author would automatically receive a small royalty--a one-time electronic debit--from anyone who accessed her work on the Web. Another piece is "transcopyright"--an emerging legal doctrine under which the author may permit readers to quote her work as long as they accessed it from her site. A third piece is a program that would let the reader see the original document side by side with documents that quote it.

Don't think for a minute that Ted Nelson has given up more radical plans for transforming cyberspace. He's got projects under way to redesign word processing and the construction of interactive media. It seems a long way from his college days of pen and paper, but it isn't. "Swarthmore is a place where ideas are honored," says Nelson. And all he ever wanted was "a magic paper that allows those ideas to be expressed and understood in their full glory."

William Saletan '87 is a senior writer for Slate ( His book on the politics of abortion rights is forthcoming from the University of California Press. Saletan serves as a member of this magazine's advisory committee.


Bulletin Home Page