NotesNudge the notes you never go back to

Linking Notes

Linking notes to each other is the defining feature of the current generation of tools, and the promise attached to it is large: a network that surfaces connections you did not see.

The mechanism buys something real. It is smaller and more specific than the promise, and it comes with a cost nobody mentions at the start.

For a workplace-oriented comparison with this kind of workflow, this guide provides additional context.

What a link actually buys

It records a connection at the moment you saw it. This is the whole value and it is genuinely significant. You noticed that this relates to that; the noticing lasted about ninety seconds; without the link it is gone and cannot be reconstructed from proximity or shared tags.

It gives retrieval a second route. Search fails when you cannot name what you are looking for. Following a link from a note you did find is a different path and it sometimes works when search does not.

And it forces a small act of thinking at capture. Deciding what this relates to is work, and the work is most of the benefit — the same reason restating in your own words matters.

For an independent external reference related to this topic, see Roam Research.

Notice that all three benefits are delivered when the link is made, not when the graph is viewed.

What the promise adds, without support

"The network will surprise you with connections."

It is plausible. It is widely reported by practitioners. There is no body of research showing that link-based note systems produce insight at a rate other arrangements do not, and claims about creativity are a different category from claims about memory — the memory research does not reach them.

And the graph view is decorative. It is the most photographed feature in this software and the least useful. A visualisation of two thousand nodes tells you that you have two thousand notes. Nobody has ever found something they needed by looking at one.

The costs

Links are a maintenance obligation. Notes get renamed, deleted, merged. Links rot. A system with three hundred links has three hundred small liabilities, and tools handle this unevenly.

Linking invites over-linking. Once the mechanism exists, connecting things is satisfying, and a note linked to fifteen others is linked to nothing in particular. The signal is the selectivity.

It raises capture cost. Deliberately, and that is often good — but if your problem is that you already do not write things down, this makes it worse.

And it produces lock-in. A heavily linked archive is much harder to move than a flat one, and tools do not last as long as archives, which is a real cost paid years later.

How to link well

State why, not just that. A bare link records that two notes are related. A line saying this contradicts the argument in X records what you actually saw, and it is the difference between a graph and a record of reasoning.

Link sparingly. Two or three deliberate connections beat fifteen automatic ones. If you find yourself linking because the interface makes it easy, you have started decorating.

Prefer links from new to old. Connecting a new note to existing ones is where the thinking happens. Going back to add links to old notes is organising, which is the fluent half.

And do not maintain the graph. Broken links are fine. Fixing them is maintenance work with no reader.

Who should bother

People writing from their notes. If the archive feeds essays, reports or research, links between related material earn their cost repeatedly.

People with a few long-running questions, where connections accumulate over years and the record of reasoning compounds.

Not people accumulating reference material. Search and titles are better and cost nothing.

And not people whose problem is that they never return. Linking is a capture-side improvement, and a beautifully connected archive is exactly as silent as a flat one.

The short version