Juno left the old failure on the first table.
The regulator housing stood open beside its record. Next to it lay the cutoff drawing, the units correction, and the original capacity estimate with the superseding figure attached. On the second table she had placed the new mixed-signature settings and the bounded shared-load result. Visitors would have to pass both to reach the review copies.
Eli arrived carrying the clean comparison sheets and stopped to look at the arrangement.
“You think they’ll say we’re hiding the failures?”
“They may say anything,” Juno said. “I want them to be able to find the failures.”
The dated revision identified exactly what changed from the public baseline. It described the material effect on contribution entry rate, the setting checked in the independent prototype, and the local calibration still required at a different site. It preserved the small test interval. It did not enlarge the successful load because a full-scale seal needed a larger number.
The independent operator checked the contribution ceiling against both instrument records. The dungeon engineer checked the material description. Sera compared the version marks so a reader could distinguish the baseline, its correction, and the new calibration revision without guessing which loose page was current.
