Keeping everything is not safe
It is natural to think that the safest thing to do with old data is to keep all of it. It is not. Every useless file has to be stored, migrated, checked and searched through, and every one makes the important files a little harder to find. Archives that keep everything end up with nothing anyone can locate.
I have written about how this happens in Kipple. This guide is the practical part.
Four questions
Lazarus archivists ask four questions of anything offered to the archive. They are simple, and answering them honestly is most of the job.
- Is it unique, or does a better copy exist somewhere else?
- Is it evidence of something: a decision, an event, a life?
- Would someone fifty years from now want it, and could they make sense of it?
- Can we keep it readable for that long, and at what cost?
Something that passes any one of them is usually worth keeping. Something that fails all four is let go.
Letting go, on the record
Letting go is part of preservation, not a failure of it, but it has to be done deliberately. When we decide not to keep something, we write down what it was, why we decided as we did and who decided. Deletion with a record is the opposite of loss: it is a decision someone can understand later, and question if they need to.
Duplicates
Most collections hold several copies of the things that matter most, in slightly different versions, with nobody sure which is the real one. Before deciding what to keep, find the duplicates, compare them and choose one authoritative copy. Keep the others only if they differ in a way that matters, and say how.
A few rooms clear
Clutter always comes back. The organizations that cope best do not win one heroic clear-out; they set aside a regular day, every quarter, for deciding what to keep. You will not win against it. You can keep one room clear, and that is enough.