NotesNudge the notes you never go back to

What to Review, What to Drop

The instinct with a large archive is to review it all, eventually, somehow. It does not work, and the size is not really the issue — but the arithmetic is worth seeing: two thousand notes at ten minutes a week is a cycle measured in years, by which time most of them have expired anyway.

Reviewing everything is functionally the same as reviewing nothing. The useful question is which small fraction earns a place.

A related workplace concept is chronemics, which offers a useful contrast with the personal knowledge-work problem discussed here.

Four kinds of note

Sort by what the note is for, not by subject.

Thinking. Written to process something. The value was delivered at the moment of writing. Does not need reviewing and never did. This is most of what most people write, and treating it as a debt is the source of most of the guilt in this subject.

Reference. Something you will look up when you need it. A command, a recipe, a phone number, a specification. Does not need reviewing — it needs to be findable. Search solves this and rotation does not.

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

Load-bearing. Something you want available without looking it up: an idea you want to think with, a principle you want to apply, a fact you need at hand. This is the only category that earns review, and it is small.

Expired. True once, no longer relevant, or superseded. Delete.

Most archives are perhaps sixty per cent thinking, thirty per cent reference, a few per cent load-bearing, and a growing sediment of expired.

Marking the small set

The load-bearing set has to be identified, and the cheapest moment is at capture.

Ask at capture: will I want this without looking it up? If yes, mark it. A tag, a folder, a star — the mechanism does not matter, the decision does.

Expect it to be small. If more than about one note in twenty is going in, you are marking things that are actually reference. Reference is fine; it just belongs in search rather than in rotation.

Add to it during review. A note that keeps proving useful gets promoted. This is how the set stays honest rather than reflecting what you thought was important two years ago.

And retire from it. Once something is genuinely part of how you think, reviewing it adds little — the benefits asymptote and the slot is better spent elsewhere.

What to do with the rest

Thinking notes: leave them. Searchable, unreviewed, and untroubling. If one resurfaces at random, fine — that is rediscovery, and it is what random resurfacing is for.

Reference: invest in findability rather than rotation. A good title beats any amount of review.

Expired: delete during review sessions rather than in a periodic purge. Purges are large, dreaded and postponed; deletions during a ten-minute session are three or four notes and cost nothing.

The deletion question

The part people resist, so here is the argument plainly.

An archive that only grows becomes less useful at a rate proportional to its size. Every note you keep that you will never use makes the ones you will use harder to find, and dilutes any random selection.

Deleting is not losing. The encoding benefit already happened. Writing it did the work whether or not the file survives.

And keeping everything is a decision too, made by default, with costs you pay continuously and never attribute.

The test that makes it easy: if this arrived at random tomorrow, would I be glad? If the honest answer is that you would skim it and move on, it is expired.

What a rotation actually looks like

For a ten-minute weekly session:

Three or four from the marked set, each with an attempted recall before opening.

One at random from everything else, for the rediscovery effect.

A decision on each — keep, promote, retire, delete.

That is fifteen or twenty notes a month from a marked set of perhaps forty, which means each returns roughly every two months. That is a real interval on a real set, which is more than an archive-wide rotation can ever offer.

The short version