A docs passage describes a converted relation as the view it stopped being

A sentence describes graphitron_spelled_table as a thing it stopped being when a gatherer stage started writing it, and it was edited by the commit that made it wrong. When this lands it says what the relation now is, so a contributor reading the passage learns the shape the tree actually ships.

docs/architecture/explanation/naming-the-row.adoc introduces graphitron_spelled_table as "a view that answers a small, general question". It is a base table a capture stage reconciles, and the sentence sits in a passage whose whole argument is the contrast between one stored answer table and a layered derivation, so the noun is doing work rather than passing by. The bullet below it, intent_bound_table "is a view one layer up", is still right, which is what makes the pair misleading: a reader takes both nouns at face value. What the passage wants now is the shape the conversion actually landed, derived and stored, which is a better version of the point it was making than the one it makes.

UnregisteredRelationTest’s class javadoc and its `TARGET constant named two different relations, the javadoc still carrying the thirteen-readers rationale of a relation that had stopped being registered while the constant had moved to one that still was. R955’s first rung converted that second relation in turn and repointed both together, so this surface reads true in the shipped tree and only the naming-the-row passage above is left. The pattern the pair showed is what outlives it: a conversion moves the constant, and a javadoc renamed in place rather than repointed states a rationale about a relation nothing in the class names any more.

No new case. The javadoc reference gate does not read prose nouns and no gate compares a class comment against a constant beside it, so this is a read-and-correct rather than something to pin.

Found at R954’s In Review → Done gate, in the tree that item shipped; both sentences were touched by its rename pass. Filed rather than corrected there, the gate being findings-not-fixes and the corrections belonging to whoever next reads the passages whole.