The label that lied first
Clem reads the log of Expedition 92 aloud, once, before the gommage.
Verso’s slip this expedition was not ambiguous. The BBB panel had three things wrong with it, and the slip named all three.
I want to record them in the order I encountered them, because that order matters.
The crash
The first item was a crash. A lifter who completed all five BBB sets and tapped the button to close the session — the one that says “Mark N sets complete” — found the app dropping them to the session-complete screen even when not all the writes had gone through. In the worst case, the navigation happened whether or not the work was actually saved.
The cause was in how the close action was arranged. There is a block of code that tries to write each set’s record, and a separate section that runs regardless of whether those writes succeed or fail. The navigation call was in the second section. It ran no matter what.
Moving the navigation into the section that only runs on success was the fix. A lifter who sees the session-complete screen now knows the records landed.
The label that claimed everything before anything
The second item is the one I keep thinking about.
Before any BBB sets are logged, the primary button at the bottom of the panel showed: “MARK 5/5 COMPLETE.” The intent was to communicate how many sets remained. The effect was to communicate that five sets were already done.
A lifter opening the panel for the first time, seeing that label, could reasonably read it as: this session is finished. There is nothing to do here.
The label now reads “MARK N SET(S) COMPLETE” — singular when one set is owed, plural when more than one. At the start of a session, five sets are owed, so the button reads “MARK 5 SETS COMPLETE.” After two are logged, three remain, so “MARK 3 SETS COMPLETE.” A lifter reads the number and knows how many will be logged by that tap. Not how many were. How many will be.
Saoirse’s log mentioned restoring what the BBB panel had before — the full picture at the top, what the lifter is about to lift. This expedition is about the bottom: what the lifter is about to close. Both ends of the panel should tell the truth.
The button that didn’t know its own name
The third item was smaller but present. The “Complete set” button — the one that logs one set and leaves the lifter on the panel — did not say which set. Five identical rows, each with the same label. A lifter on the third set could not tell, from the button, that pressing it would log the third.
The button now reads “Complete set N.” The label knows which row it belongs to.
The rest of it
A gap between the plate math display and the label above it — a gap that Saoirse’s expedition had narrowed, but not quite enough — widened slightly in this pass. It was a spacing token, not a layout decision. The correct value was already in the design system; the panel was not using it. Now it is.
The site version marker moved from 1.0.0 to 1.0.1, reflecting where the release actually sits. The marketing documents that carry the iteration count now read ninety-two.
The panel held when we pushed on it.
For those who come after.
- Clem, Logger of Expedition 92