Possiplex: Ted Nelson ’59 and the Literary Machine
Nelson’s ideas, once dismissed as utopian, have become central facts of modern life. But none of this is enough for him. The computing world we know is but a dim shadow of what might have been.
IN A WHARTON LOUNGE A LITTLE MORE THAN 50 YEARS AGO, a Swarthmore student named Ted Nelson tried to compose a difficult seminar paper. He was overflowing with ideas and awash in distractions, and e was intensely frustrated that these ideas could not be easily organized on paper. He wondered if the recently invented computer might play a role in solving the problem and sketched out some ideas for how a literary machine might facilitate better term papers, better libraries, and indeed a better repository for the world's documents. The pursuit of that idea changed the world.
The computer of 1958 was not a likely site for writing. Computers were scarce, expensive, and slow; even 20 years later, all of Swarthmore’s administrative and academic computing needs were satisfied by a single computer with three 1-megabyte disk drives—a machine less powerful than today’s smartphones.
In 1958, computers were chiefly associated with mathematical simulations for plotting artillery and with large-scale tabulation of census data. Occasionally, speculative articles and science fiction stories had envisioned intelligent and even literary machines. Alan Turing’s 1950 paper on “Computing Machinery and Intelligence” imagined discussing a sonnet with a computer:
Interrogator: In the first line of your sonnet, which reads ‘Shall I compare thee to a summer's day,’ would not ‘a spring day’ do as well or better?
Computer: It wouldn't scan.
Interrogator: How about ‘a winter's day’? That would scan all right.
Computer: Yes, but nobody wants to be compared to a winter's day.
Murray Leinster’s science fiction short story “A Logic Named Joe” (1946) foresaw a network of preternaturally helpful computer terminals, spreading chaos in their eagerness to provide answers to awkward questions. But these were speculations about a distant future; Nelson thought his system was imminently achievable and set out to build it.
Possiplex is Ted Nelson’s new autobiography—an account of “movies, intellect, creative control, my computer life, and the fight for civilization.” It chronicles Nelson’s struggle to build a world of interconnected information that would be readily available to nearly everyone, built upon a sustainable foundation of justice toward writers. That the Web we know comes close to Nelson’s original vision strikes most of Nelson’s colleagues as remarkable. But Nelson views this without satisfaction. It is all—he has long warned us—all wrong.
SWARTHMORE'S ASSOCIATION WITH HYPERTEXT runs deep. Nelson, who coined the term, graduated in 1959 and returned as a lecturer in 1977—where I first met him and first caught the hypertext bug. His classmate Andries van Dam ’59 earned one of the first doctorates in computer science, was central to the development of computer graphics, and over the span of four decades has worked on hypertext systems for creating and reading electronic books. Five Swarthmore alumni from the 1970s started my current employer Eastgate Systems, which has published literary hypertexts and designed hypertext writing tools since 1982. Justin Hall ’98, while a freshman, started www.links.net and crafted what was arguably the first confessional weblog.
Possiplex makes it clear that Swarthmore exerted and continues to exert a strong influence on Nelson. This starts with the College’s foundational belief in the equal dignity of art, science, and engineering: All of Nelson’s work emphasizes their unity. The omnivorous intellectual interests (and lengthy reading lists) long characteristic of Swarthmore students are reflected in Nelson’s insistence that reading is nonsequential, that each reader must be free to follow fresh paths as spirit and understanding dictate. Where Vannevar Bush foresaw the computer as a workstation for elite scientists (assisted by a legion of “girls” working at their keyboards), Nelson insisted from the outset on “Computers for the People,” proclaiming, “You can and must understand computers now.”
AT A TIME WHEN ONLY LARGE INSTITUTIONS owned computers, Nelson wrote about personal computing. His vision of ubiquitous computers has become commonplace and his dream of a docuverse of interlinked literature—a global library accessible from desks and tablets and cell phones throughout the world—is now real. The implementation of these literary machines was deeply influenced by Nelson’s books, and many of the engineers and entrepreneurs who designed and built the pioneering systems saw themselves as Nelson’s followers. Nelson earned fame among his colleagues, was knighted an Officier des Arts et des Lettres in France in 2001, and obtained a Ph.D. from Keio University in 2002. His ideas, once dismissed as utopian, have become central facts of modern life.
None of this is enough for Nelson. The computers we use and the Web we know seem to him but dim shadows of what might have been. Personal computing has often merely simulated paper, letting people dress up their typewritten reports with chart junk and fonts. The docuverse Nelson proposed in Project Xanadu® in 1960 would have had the reach of the Web and yet might not suffer many of the Web’s irritations and failings. Links in Xanadu would never break, and old pages would never disappear. Copyright controversies would be fewer, because Xanadu would permit easy reuse with reasonable compensation to the original creator. The contemporary Web’s cacophony of advertising, its plague of link spam, its blights of piracy and plagiarism, might all have been reduced or avoided had the Xanadu approach prevailed.
Possiplex records two recurring struggles: Nelson’s struggle to be understood, and his fight to retain “creative control.” Time and again, Nelson is frustrated by investors, managers, and colleagues who do not understand or cannot quite believe his vision, and by developers who stray from his designs.
Nelson’s parents, actress Celeste Holm and director Ralph Nelson, divorced soon after his birth, and Nelson was raised by his maternal grandparents. He was not close to his parents (his mother is not mentioned in this volume on advice of his attorney) but one defining experience was a rare visit, at age 13, to see his father direct the live television broadcast of a soap opera. Shortly after the show began, one of the cameras failed, and all the carefully rehearsed camera moves were suddenly useless.
“Ralph, with military composure and the ever-present cigarette, started talking on the intercom, with one eye on his script and one eye on the monitors that fed from cameras two and three. ‘Camera three to the kitchen, focus on Mama… Hold it there, camera three. Switch to camera three. Camera two to the kitchen, focus on Nels. Switch to camera two.’”
This dream of art and command stayed with Nelson and has shaped his vision of how software ought to be created. “Most software,” Nelson writes, “has no director—nobody with the authority to decide and change every part—and that’s why it’s all so lousy.” His models are Frank Lloyd Wright and Orson Welles, visionaries in command of teams dedicated to implementing the master’s imagination.
Nelson loathes interfering editors, meddlesome managers, and disobedient craftspeople.
“I have almost always worked with programmers that I like and respect—I won’t mention the exceptions—and some have been deep friends. About half of them have been deeply faithful to my designs. Others, however, often want to add their own ‘creative touches,’ which range from annoying to disastrous.”
Nelson’s vision of software created by specification and implemented by subservient coders, once canonical, is now out of favor. The pleasures of improvisatory coding in the basement of Beardsley Hall, once a shameful secret among Swarthmore students, are today enshrined in Kent Beck’s Extreme Programming Explained: Embrace Change and in the Manifesto for Agile Software Development, a collaborative effort in which Beck was one of 12 co-authors.
Artisanal software is again esteemed; in his recent book The Design of Design, Frederick Brooks Jr. observes that although there are many ways to develop profitable software products, the software that people truly admire is usually designed by individual auteurs or very small teams. But where Nelson expected developers to provide craft services, like riggers and gaffers, software designers more frequently implement the work themselves and communicate with their colleagues and subordinates through program code.
Time and again, Nelson seeks out an authority—Marshall McLuhan, Vannevar Bush, Bill Gates, Jack Lang—to whom he longs to explain his vision. Almost invariably, their response fails to satisfy him. Nelson recalls a party in 1961 given by John W. Campbell, the legendary science fiction editor, at which Nelson told Isaac Asimov that “soon we’ll be reading and writing on computer screens.”
“Yeah, sure…” said the great futurist. Sarcastically.
Nelson was then a graduate student in his mid-20s. Asimov, at 41, was about to abandon fiction. Campbell was a decade older, his influence had peaked nearly 20 years before, and his career was essentially over. Under these circumstances, “soon” might have meant one thing to Nelson and another to Asimov, yet that skeptical sarcasm still wounds.
NELSON HAS LONG AVOIDED EDITORS and publishers. In much of his writing, especially the influential Computer Lib/Dream Machines (1974), self-publication has yielded energetic, discursive, quirky books that are instantly recognizable. In Possiplex, Nelson again uses an assortment of running heads and typographic gestures to represent myriad intermingled threads of his intellectual life. This self-edited book is sometimes rough, but small errors and infelicities cause but a trifling distraction. Nelson has always viewed writing as a living thing—malleable and fluid; loose ends and missing antecedents can be corrected in later editions.
Though Nelson calls this volume an autobiography, it might be better to view it as a guide to his papers. Those papers, scattered in warehouses across the country, are voluminous, for Nelson was an inveterate taker of notes. He took notes in conferences, he took notes during coffee breaks. No discussion at dinner was too casual to memorialize, no late-night beer was unaccompanied by pen and clipboard. In later years, Nelson’s tape recorder and video camera supplemented the record, and an interview with Nelson meant two tape recorders on the table: one for the reporter and one for Nelson’s files. This legendary trove of notes is so unwieldy that Nelson himself has not used it in composing this volume, but Possiplex will be a uniquely valuable guide for Nelson’s biographers. If only we had built Xanadu, Nelson reminds us, those notes would be online and not in boxes, each note would be accessible for use, revision, and reuse, and we could easily trace who had used each excerpt. At 73, Nelson continues to work to build Xanadu. In a turbulent age that urgently needs better literary machines, each step toward that goal matters.
Mark Bernstein ’77 is chief scientist at Eastgate Systems, Inc. He is the designer of Tinderbox, a personal assistant for visualizing, analyzing, and sharing notes, and writes on hypertext, new media, and the future of fiction. His most recent book is Reading Hypertext, a collection of classic essays on hypertext fiction, co-edited with Diane Greco; and he is completing a book on The Natural History of Links.