💾 Archived View for tilde.club › ~winter › gemlog › 2023 › 4-05.gmi captured on 2024-07-09 at 01:48:58. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2024-05-26)

-=-=-=-=-=-=-

Daniel Janus, Ted Nelson, and the Web of Documents

A couple of weeks ago I read a blog entry by Daniel Janus on what he calls the web of documents, tracing the idea of a digital document repository from Ted Nelson's Xanadu project to the current state of the World Wide Web.

The Web of Documents

Project Xanadu (official)

Project Xanadu (The History of Computing)

Xanadu was, or is, an extremely ambitious project that aimed to build a durable, world-wide network of documents. At first glance, that seems to be the original premise of the web, or more recently Gemini - the promise of document retrieval with internal links to documents stored elsewhere. But the resemblance is only superficial: one of the oldest problems of the web is the broken link, whereby a resource at a particular address no longer exists there, possibly moving elsewhere. If the site owner is thinking ahead, there will be HTTP redirects. But these are by no means mandatory, and often forgotten. The longer a page stays online, the greater the chance its links begin to rot.

Project Xanadu (Wikipedia)

Ted Nelson started Project Xanadu in 1960, when he was a graduate student at Harvard. He envisioned a network of documents existing purely digitally. In those documents were paths to other documents - he called this hypertext - and they could additionally contain portions of other documents. Nelson called this transclusion. Others also saw the potential of the electronic, interactive computer with respect to documents: Douglas Englebart included hypertext as part of his groundbreaking Mother of All Demos in 1968.

The Mother of All Demos

The Curse of Xanadu

Orality and Hypertext: An Interview with Ted Nelson

Fairly or not, Ted Nelson is most remembered today for the grandness of his Xanadu, and how he never fully delivered on its promise. Indeed, a "working deliverable" of Xanadu took until 2014 to be released. By comparison, the Macintosh shipped with HyperCard in 1987; the World Wide Web took off in the early 90s.

There is a note of bitterness in Ted Nelson when he writes about the web, the greatest and worst realization of his vision:

Ted Nelson's Computer Paradigm, Expressed as One-Liners

The Web is a foam of ever-popping bubbles, ever-changing shopwindows.
HTML is precisely what we were trying to PREVENT-- ever-breaking links, links going outward only, quotes you can't follow to their origins, no version management, no rights management.

And so on. But what Nelson, and Janus, miss in their evaluation of the web is that the creation of a particular technology is a starting point, not a promise. Berners-Lee et al. saw the complexities associated with a full implementation of a Xanadu like system, and charged past them, implementing a working minimal subset. They authored standards and protocols and formed working groups to allow future standards to move ahead. What could you do in the early days? Documents could be retrieved and they could be marked up by the creators using the then-simple HTML. Links broke, but you could email someone and they might update them. It wasn't perfect, but it worked. It caught on. It let people do things in fundamentally different ways. People had ideas what to do next.

It was good enough.

After the early years, after the web escaped its academic confines and spread across the world, it transformed. People still put "documents" online, but they also added music, or movies, or applets that let them play games and chat. Eventually this transformed into the backbone of the app ecosystem of the smartphone age, our phones talking to remote servers using HTTPS.

Janus identifies this shift as a transformation from a web of documents to a web of applications. But the reason for this shift sits at the end of each connection, as we wait by our computers and phones. The web has always ultimately been a web of people, not applications. And though that feels less true today, as the protocol serves fewer actual pages and more API responses, it's become an indispensible part of our lives. We've gone from those first few minutes online, to days, years, decades.

It's not great but it's good enough. It solves a problem and leaves room for improvement. I think about this a lot, especially with respect to Gemini, which attempts to solve a particular problem of the web (co-opting protocols and processes to undermine the original vision of the medium) by operating off a closed and unextendable protocol. Maybe that's essential to prevent a repeat of what we've seen with the web. But the lesson I've taken from the web and Xanadu is that a pure or better system is useless without users. And if you're not making something that captures people's imaginations, you'll spend years complaining bitterly about the people who did.

I don't think many people expect Gemini to take over the world. Likely it will keep and grow its current niche, which seems to be a cross-section of the technologically-inclined. But Gemini went from an idea to a protocol to a set of implementations. It's nowhere near as popular as the web, but it's got more to it than Xanadu. I navigate through unfamiliar capsules in Lagrange. I proofread my gemlog entries in amfora. It's not perfect, but it works, there's things I'd change, but it's good enough, and crucially, other people feel the same. People are still signing up for accounts on the many entry points, or registering domains and putting up servers themselves. They're building something fragile and durable, making capsules, creating gemlogs and gemrolls and link directories, repeating what Ted Nelson might consider to be the same mistakes of the web, but happily, willingly, because the heart of this system is the people using it. And they're online every day, typing into unstyled text areas, or logged into tildes putting up pages for strangers, and making their world a little bigger.

gemlog