May 13, 2021
Java, Ruby, PHP, Go. These are web applications that dynamically generate code then interpreted as a file by a web browser. That file is rarely static these days and the power of the web is that an app or browser can reach out and obtain some data, get back some xml or json or yaml, and provide an experience to a computer, mobile device, or even embedded system. The web is arguably the most powerful, transformational technology in the history of technology.
But the story of the web begins in philosophies that far predate its inception. It goes back to a file, which we can think of as a document, on a computer that another computer reaches out to and interprets. A file comprised of hypertext. Ted Nelson coined the term hypertext. Plenty of others put the concepts of linking objects into the mainstream of computing. But he coined the term that he’s barely connected to in the minds of many. Why is that?
Tim Berners-Lee invented the World Wide Web in 1989. Elizabeth Feinler developed a registry of names that would evolve into DNS so we could find computers online and so access those web sites without typing in impossible to remember numbers. Bob Kahn and Leonard Kleinrock were instrumental in the Internet Protocol, which allowed all those computers to be connected together, providing the schemes for those numbers. Some will know these names; most will not.
But a name that probably doesn’t come up enough is Ted Nelson. His tale is one of brilliance and the early days of computing and the spread of BASIC and an urge to do more. It’s a tale of the hacker ethic. And yet, it’s also a tale of irreverence - to be used as a warning for those with aspirations to be remembered for something great. Or is it?
Steve Jobs famously said “real artists ship.” Ted Nelson did ship. Until he didn’t. Let’s go all the way back to 1960, when he started Project Xanadu. Actually, let’s go a little further back first.
Nelson was born to TV directory Ralph Nelson and Celeste Holm, who won an Academy Award for her role in Gentleman’s Agreement in 1947 and took home another pair of nominations through her career, and for being the original Ado Annie in Oklahoma. His dad worked on The Twilight Zone - so of course he majored in philosophy at Swarthmore College and then went off to the University of Chicago and then Harvard for graduate school, taking a stab at film after he graduated. But he was meant for an industry that didn’t exist yet but would some day eclipse the film industry: software.
While in school he got exposed to computers and started to think about this idea of a repository of all the world’s knowledge. And it’s easy to imagine a group of computing aficionados sitting in a drum circle, smoking whatever they were smoking, and having their minds blown by that very concept. And yet, it’s hard to imagine anyone in that context doing much more. And yet he did.
Nelson created Project Xanadu in 1960. As we’ll cover, he did a lot of projects during the remainder of his career. The Journey is what is so important, even if we never get to the destination. Because sometimes we influence the people who get there. And the history of technology is as much about failed or incomplete evolutions as it is about those that become ubiquitous.
It began with a project while he was enrolled in Harvard grad school. Other word processors were at the dawn of their existence. But he began thinking through and influencing how they would handle information storage and retrieval.
Xanadu was supposed to be a computer network that connected humans to one another. It was supposed to be simple and a scheme for world-wide electronic publishing. Unlike the web, which would come nearly three decades later, it was supposed to be bilateral, with broken links self-repairing, much as nodes on the ARPAnet did. His initial proposal was a program in machine language that could store and display documents. Being before the advent of Markdown, ePub, XML, PDF, RTF, or any of the other common open formats we use today, it was rudimentary and would evolve over time. Keep in mind. It was for documents and as Nelson would say later, the web - which began as a document tool, was a fork of the project.
The term Xanadu was borrowed from Samuel Taylor Coleridge’s Kubla Khan, itself written after some opium fueled dreams about a garden in Kublai Khan’s Shangdu, or Xanadu.In his biography, Coleridge explained the rivers in the poem supply “a natural connection to the parts and unity to the whole” and he said a “stream, traced from its source in the hills among the yellow-red moss and conical glass-shaped tufts of bent, to the first break or fall, where its drops become audible, and it begins to form a channel.”
Connecting all the things was the goal and so Xanadu was the name. He gave a talk and presented a paper called “A File Structure for the Complex, the Changing and the Indeterminate” at the Association for Computing Machinery in 1965 that laid out his vision. This was the dawn of interactivity in computing. Digital Equipment had launched just a few years earlier and brought the PDP-8 to market that same year. The smell of change was in the air and Nelson was right there.
After that, he started to see all these developments around the world. He worked on a project at Brown University to develop a word processor with many of his ideas in it. But the output of that project, as with most word processors since - was to get things printed. He believed content was meant to be created and live its entire lifecycle in the digital form. This would provide perfect forward and reverse citations, text enrichment, and change management. And maybe if we all stand on the shoulders of giants, it would allow us the ability to avoid rewriting or paraphrasing the works of others to include them in own own writings. We could do more without that tedious regurgitation.
He furthered his counter-culture credentials by going to Woodstock in 1969. Probably not for that reason, but it happened nonetheless. And he traveled and worked with more and more people and companies, learning and engaging and enriching his ideas. And then he shared them.
Computer Lib/Dream Machines was a paperback book. Or two. It had a cover on each side. Originally published in 1974, it was one of the most important texts of the computer revolution. Steven Levy called it an epic. It’s rare to find it for less than a hundred bucks on eBay at this point because of how influential it was and what an amazing snapshot in time it represents.
Xanadu was to be a hypertext publishing system in the form of Xanadocs, or files that could be linked to from other files. A Xanadoc used Xanalinks to embed content from other documents into a given document. These spans of text would become transclusions and change in the document that included the content when they changed in the live document. The iterations towards working code were slow and the years ticked by. That talk in 1965 gave way to the 1970s, then 80s. Some thought him brilliant. Others didn’t know what to make of it all. But many knew of his ideas for hypertext and once known it became deterministic.
Byte Magazine published many of his thoughts in 1988 called “Managing Immense Storage” and by then the personal computer revolution had come in full force. Tim Berners-Lee put the first node of the World Wide Web online the next year, using a protocol they called Hypertext Transfer Protocol, or http. Yes, the hypertext philosophy was almost a means of paying homage to the hard work and deep thinking Nelson had put in over the decades. But not everyone saw it as though Nelson had made great contributions to computing.
“The Curse of Xanadu” was an article published in Wired Magazine in 1995. In the article, the author points out the fact that the web had come along using many of the ideas Nelson and his teams had worked on over the years but actually shipped - whereas Nelson hadn’t. Once shipped, the web rose in popularity becoming the ubiquitous technology it is today. The article looked at Xanadu as vaporware. But there is a deeper, much more important meaning to Xanadu in the history of computing.
Perhaps inspired by the Wired article, the group released an incomplete version of Xanadu in 1998. But by then, other formats - including PDF which was invented in 1993 and .doc for Microsoft Word, were the primary mechanisms we stored documents and first gopher and then the web were spreading to interconnect humans with content.
The Xanadu story isn’t a tragedy. Would we have had hypertext as a part of Douglas Engelbart’s oNLine System without it? Would we have object-oriented programming or later the World Wide Web without it? The very word hypertext is almost an homage, even if they don’t know it, to Nelson’s work. And the look and feel of his work lives on in places like GitHub, whether directly influenced or not, where we can see changes in code side-by-side with actual production code, changes that are stored and perhaps rolled back forever.
Larry Tessler coined the term Cut and Paste. While Nelson calls him a friend in Werner Herzog’s Lo and Behold, Reveries of the Connected World, he also points out that Tessler’s term is flawed. And I think this is where we as technologists have to sometimes trim down our expectations of how fast evolutions occur. We take tiny steps because as humans we can’t keep pace with the rapid rate of technological change. We can look back and see a two steps forward and one step back approach since the dawn of written history. Nelson still doesn’t think the metaphors that harken back to paper have any place in the online written word.
Here’s another important trend in the history of computing. As we’ve transitioned to more and more content living online exclusively, the content has become diluted. One publisher I wrote online pieces for asked that they all be +/- 700 words and asked that paragraphs be no more than 4 sentences long (preferably 3) and the sentences should be written at about a 5th or 6th grade level. Maybe Nelson would claim that this de-evolution of writing is due to search engine optimization gamifying the entirety of human knowledge and that a tool like Xanadu would have been the fix. After all, if we could borrow the great works of others we wouldn’t have to paraphrase them. But I think as with most things, it’s much more nuanced than that.
Our always online, always connected brains can only accept smaller snippets. So that’s what we gravitate towards. Actually, we have plenty of capacity for whatever we actually choose to immerse ourselves into. But we have more options than ever before and we of course immerse ourselves into video games or other less literary pursuits. Or are they more literary? Some generations thought books to be dangerous. As do all oppressors. So who am I to judge where people choose to acquire knowledge or what kind they indulge themselves in. Knowledge is power and I’m just happy they have it. And they have it in part because others were willing to water own the concepts to ship a product. Because the history of technology is about evolutions, not revolutions. And those often take generations. And Nelson is responsible for some of the evolutions that brought us the ht in http or html. And for that we are truly grateful!
As with the great journey from Lord of the Rings, rarely is greatness found alone. The Xanadu adventuring party included Cal Daniels, Roger Gregory, Mark Miller, Stuart Greene, Dean Tribble, Ravi Pandya, became a part of Autodesk in the 80s, got rewritten in Smalltalk, was considered a rival to the web, but really is more of an evolutionary step on that journey. If anything it’s a divergence then convergence to and from Vannevar Bush’s Memex.
So let me ask this as a parting thought? Are the places you are not willing to sacrifice any of your core designs or beliefs worth the price being paid? Are they worth someone else ending up with a place in the history books where (like with this podcast) we oversimplify complex topics to make them digestible? Sometimes it’s worth it. In no way am I in a place to judge the choices of others. Only history can really do that - but when it happens it’s usually an oversimplification anyways… So the building blocks of the web lie in irreverence - in hypertext. And while some grew out of irreverence and diluted their vision after an event like Woodstock, others like Nelson and his friend Douglas Englebart forged on. And their visions didn’t come with commercial success. But as an integral building block to the modern connected world today they represent as great a mind as practically anyone else in computing.