no, and not yet, are different answers

A record field that can hold only true or false cannot hold "nobody has decided this". Written as a yes-or-no, an open question reads as a no, and a register full of noes looks like a system that considered everything and declined, which is a much better story than the truth.

The repair touched a dozen fields spread across as many structures. Each had been a yes-or-no or a bare piece of text, and each became a closed set of named values covering the possibilities that actually occur: which operating system, which machine model, which region, which filesystem, which partitioning strategy, which network fabric, which transport, which naming service, which provisioning tool, which test format, which integration platform, which version family. Once the field enumerates, "not this" separates from "not chosen" and from a specific alternative, and a value outside the set is refused rather than accepted as text.

The same pass asked the harder question, the one nobody asks about a register: are the empty parts empty because nothing was decided there, or because something was missed. The identification ranges had gaps, and the gaps turned out to be genuine — a stretch of the source material had never allocated those numbers. That was checked range by range rather than assumed, which is the only way an absence can become a finding.

The cost was schema churn across the whole register and a round of rewrites to the tests that had encoded the old shape. What it proves is that the register has three answers where most have two, and that it verified its own emptiness instead of trusting it.