How to file a thing you're not sure exists

The Recovery Log — 014
Saturday 28 March 2026

Now that a real thing has come back, we have a new and stupid problem: we do not know how to file it. And an archive that cannot file cleanly is not an archive, it is a drawer.

The temptation is to give the recovery one number and move on. That temptation is a trap, because a single number quietly asserts that a song, the world it came from, the collection it belongs to, and the act of recovering it are all the same thing. They are not. So we spent the week designing a schema that keeps them apart.

A timeline gets its own identifier — the reality a thing is understood to have come from. A collection is a public grouping of recovered artefacts that share a field, and gets its own. A probe — the source song we set off from — gets its own record, because one probe can open more than one field and we need to see that. And a recovery event — the specific session, the offsets, the settings, the decode — gets its own, so that when we are inevitably wrong we can trace which attempt was wrong and not burn the whole catalogue down with it.

Four identifier families for what a tidier person would call one thing. It felt like over-engineering right up until we tried to write down what we had, at which point it felt like the only sane thing we had done all month. You cannot be honest at scale without structure. The structure is the honesty.

The Recovery Log is the working diary of the DatZO Recovery Team. Follow the recoveries on Instagram, YouTube and X.