Screens, Research and Hypertext

Powered by 🌱Roam Garden

Project Xanadu

Visionary hypertext system or long-running vaporware program?

Project Xanadu is Ted Nelson's dream of an alternative system of hypertext. Conceived nearly 30 years before the advent of the web, Xanadu failed to get off the ground (though a somewhat limited version of the system did eventually launch).

For Nelson, hypertext was a way of fully documenting the connections between ideas—of understanding that all of human knowledge is intertwingled. In many ways, Xanadu was to be a computer network that mimicked Nelson's own brain.

Screenshot of a working Xanadu, showing multiple documents with pointers demonstrating where each text is linked to others.

Alas, Nelson is more philosopher and sociologist than engineer. His vision was over-complex—perhaps flatly impossible, given the limitations of computers and computer networks at the time. And it relied on recruiting engineers to the cause, and finding the funding to pay them.

It was no small task those engineers faced, either. Xanadu wouldn't work with existing documents, data storage, or OS interfaces. It needed an entire infrastructure around it to work properly.

It's little surprise that a much simpler hypertext system—one that Tim Berners-Lee could code over Christmas and present at a conference the following January—became synonomous with hypertext.

Yet many of Xanadu's core features are objectively better than the web as we know it. And no small number could be implemented now. Maggie Appleton describes these as "the pattern language of project Xanadu." They include:

Transpointing windows. A combination of patterns one and two, transpointing windows visibly show the connections between any two given documents. Azien Elza mocked up an example using Roam Research.

Stable, universal addresses. Nelson continues to rail at the idea that we address online documents based on server locations. He describes URLs as "one-way, ever-breaking links, a shop window model," and says that it's akin to "saying that such and such a book is the book you'll find on the fourth shelf, third from the right." It seems normal enough these days to point to a URL to find content. But consider: we don't do that for email, for all that it also resides at a specific location on a server somewhere. There is no reason we couldn't treat other digital content that way rather than by pointing to server locations.

Multiple views and spatial arrangements. Xanadu would feature more than one way to visualize and navigate content, including a graph overview that is standard in Roam—available by clicking the little diagram icon up in the top right corner of your screen—and in Twine, which is where I drafted the first version of this project.

Micropayments. Authors would be compensated for reading and/or reusing their text. It's worth noting that dealing with copyright on the web remains one of its largest unsolved problems.

Nelson understood that hypertext was more than just a tool for publishing and sharing documents electronically. It was a way to extend the human brain—to build the "enlarged intimate supplement to memory" that Vannevar Bush dreamed of.

It's a system tailored for conducting research.

For more context

The web wasn't the first hypertext system. Just the most successful.

What to read next

The hypertext of Nelson's dreams models the way scholars read.

Other items of interest

Many of the design patterns of Xanadu would make the web a better place for research.

Nelson is right. Reading research texts on the web is miserable.

Referenced in

Bidirectional Links

Up until the advent of the web, pretty much everyone who talked about hypertext assumed that links would be bidirectional. Indeed, plenty of hypertext theorists are still mad about mono-directional links.

The Web as It Should Have Been

Nothing on this list is imaginary. This is no Project Xanadu-style wishful thinking. Indeed, most of the technologies needed to make each of these things happen are older than the web.


The idea of transclusion predates the web by nearly a decade. Ted Nelson coined the term in Literary Machines, his 1981 articulation of the principles behind Project Xanadu. Among those principles:

A Brief History of Hypertext

Ted Nelson coins the term "hypertext" as part of his proposed Project Xanadu, an ambitious hypertext system that would allow readers to create “zippered lists” that form compound documents from various pieces of other documents. Unlike today’s hyperlink, Xanadu’s links were bidirectional.