Research note
Semantic organization vs folder organization
It's tempting to treat automatic organisation as strictly better than folders. It isn't. The two models trade different things, and the trade only favours semantics under specific conditions.
Those conditions happen to describe most personal libraries.
01 · The case for folders
Deterministic, inspectable, portable.
A folder does what you told it to. You can see the whole structure, reason about it, back it up, hand it to someone else and open it in any tool ever made. Nothing about it is probabilistic.
For material with genuinely stable categories, tax years, client deliverables, legal records, that determinism is worth far more than any clever grouping.
02 · The case for semantics
Multiple membership, retroactive, and free.
Meaning-based grouping lets an item belong everywhere it belongs, reorganises itself when your understanding changes, and costs no attention at save time. It also works on material you'd never have filed at all, which in most libraries is the majority.
The catch: it's inferential. Groups can be wrong, and their boundaries are fuzzier than a directory's.
- Folders win on
- Certainty, portability, legal and archival material, shared team structure.
- Semantics win on
- Volume, ambiguity, multi-project items, anything never filed.
- Both need
- Search that works when the arrangement fails you.
03 · In practice
Use folders where you have certainty, semantics everywhere else.
The realistic arrangement is layered: a small, deliberate folder structure for the material that deserves one, and an automatic layer over everything else, which is where the mess actually lives.
The mistake isn't using folders. It's expecting a hierarchy to carry a library it was never shaped for.
Structure the certain things. Let the rest be found.
Next thought
Even perfect organisation doesn't help if the pile keeps growing unread.
Read: why saving more doesn't helpIf 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.