Research note
Why bookmarks failed
Bookmarks aren't broken. They do exactly what they were designed to do in 1993: record an address so you can return to it. The failure is that we started using them as a memory system.
Four properties, each fine in isolation, fatal in combination.
01 · Mechanism
Pointer, no content, no context, no decay.
A bookmark holds a URL and a title. It doesn't hold what the page said, why you saved it, or what you were doing at the time. It is a pointer to a document that may itself change or vanish.
It's also permanent by default and ordered by save time, so the list only grows and its default sort is the one attribute least connected to relevance.
- Pointer only
- Nothing searchable travels with the save.
- Title as description
- Whatever the page's own tag happened to say.
- Save-time ordering
- Sorted by the least meaningful attribute available.
- No decay, no return
- Nothing ages out, and nothing comes back to you.
02 · Scale
The failure is entirely a function of volume.
At thirty bookmarks the bar is a good tool. At three thousand it's an unreadable list with a search box that matches only titles. Nothing changed except quantity.
Folders were the offered remedy, and they inherited every hierarchy problem: single home, early guess, ongoing upkeep. Nested bookmark folders may be the least-maintained structure in personal computing.
A system that only accumulates will eventually only accumulate.
03 · Successors
Read-it-later apps improved the storage, not the memory.
The generation that followed added saved content, a reader view, tags and cross-device sync, all genuine improvements. The unread queue still grew without limit, because none of them changed the retrieval model: you still had to remember that something was in there and go looking with a keyword.
The unsolved half was always the way back in.
Next thought
The same analysis applies to the folders you file everything else into.
Read: folders and human memoryIf this idea interested you…
Shelf V3 is out now
Shelf V3 is the current build. Shelf does not use automatic updates, so download the latest build here and install it over your current version.
This is intentional. Shelf is built around a local-first, privacy-focused philosophy. An automatic updater requires additional background processes, network connections, update infrastructure, and another system with permission to modify the app. We prefer to keep Shelf’s network footprint and background activity as small as possible.
Manual updates keep the process explicit: Shelf only checks the internet when a feature actually needs it, and you decide when a new version is downloaded and installed.