A version of this post first appeared in my newsletter. Subscribe to receive posts like this in your inbox every other Sunday.

We’re used to the web as it exists today: content pages and applications you call up with a web browser. These pages and applications are published on servers by individuals or organizations responsible for their content and functionality. The operation is primarily one-way.

The original intent was slightly different; Tim Berners-Lee conceived of the web as a document management system for scientific collaboration. The first browser, WorldWideWeb, had an editing mode: you were meant not just to read documents published by others but also to write your own. The result would be a network of linked ideas you could explore and add to, a living thing that would grow and become more useful over time.

At the core of this vision is the hyperlink: a way to quickly and easily jump from one document to another (or to a specific part of a document.) Hyperlinks existed well before Berners-Lee started his project; the web was just the first instance to be adopted widely. Its success is a testament to the power of the idea coupled with a simple, open protocol.

But today, vastly more people read than publish on the web. And someone could conceivably use the web for years without realizing that linking is something they can do. The system’s client-server structure discourages users from thinking of themselves as authors.

Next-generation note-taking tools such as Roam and Obsidian harken back to the web’s original vision as a hypertext document management system. Like the web itself, links make these tools different. They encourage the creation of independent documents that are strung together in ad hoc clusters.

Many previous digital note-taking apps also support links. For example, Apple’s Notes app lets you create links in a note. However, Notes.app assumes you want to link to things outside the app, such as web pages. It doesn’t make it easy to link from one note within the app to another.

What’s new with tools like Roam and Obsidian is ease-of-use: they remove most of the friction in linking. By using a very simple text protocol, they let you create inline links to other notes or create new notes that become linked automatically to whatever you’re working on, without having to think too much about it.

The ability to quickly and easily link notes within the app changes how you use it and how you think. Rather than write down all ideas related to a particular topic in a single note, you can create individual notes for each concept and string them together using links.

The main advantage of this approach is that individual ideas can inform more than one topic. Notes can be part of different networks or threads, which you, as the author, can compose on the fly. Ideas suggest other ideas, which you capture in the system as well, leading to several different clusters that share some ideas in common.

More alluringly, you can point AI agents to trawl through your notes to find patterns between ideas that you may have missed. That’s how I use Obsidian with DEVONthink: the former is how I capture granular notes, and the latter lets me spot possible relationships between them.

The results can be quite powerful. But it works best if you capture ideas granularly — i.e., one note for one idea.

We’re not used to thinking this way. Most of us were taught to take notes linearly. For example, when listening to a lecture, we write down ideas in the sequence we hear them. The result is one or more (long) pages with a bunch of ideas presented in a fixed sequence. We don’t think to start a new note with every new idea we hear in the lecture.

This isn’t intuitive. Karl Fast and I delivered our first workshop on this subject a couple of weeks ago, and one of the biggest concerns we heard from participants was the challenge of taking notes in this more granular way. It seems most of us are deeply vested in the linear “all-in-one-go” approach to note-taking. But to create effective hypertext repositories, New idea = new note is a muscle we must flex.

And these apps make it easy: you can type out a longer note as you would in the past and later extract individual ideas into separate notes. (Roam makes it even easier: it treats each paragraph as an individually addressable block you can link to from elsewhere.) You can also transclude notes within notes, so you can have the best of both worlds.

Like the original web, the intent is to create a repository of ideas you can use and grow over time. You’re reader and author — but it’s a different type of authoring that asks you to shed assumptions about what a document is. One idea, one note.

The process is still collaborative, but the person you’re collaborating with is a future version of you. Years from now, you’ll serendipitously revisit an interesting idea and consider the many paths that converged on it — and you’ll be thankful you had the prescience to start linking granular notes.