Charles JonesOnce something is filed — a decision, a defect, a published entry — it is never edited away. A correction is appended next to it, dated, and the later entry wins.
The first state of a record is structurally the oldest. The last is current. Both stay visible.
Why deletion is the wrong cure
Deleting a defective record destroys the evidence that the defect happened. What remains is a clean-looking file with no trace of how it got that way, and the next person has no reason to suspect the class of error can occur at all.
Keeping the wrong version, clearly marked, is what lets somebody later ask how did this get here and receive an answer.
The same rule governs the incident log
Entries are appended, never rewritten. A correction becomes the next entry rather than an edit to an old one, and entries that are still open are stamped open rather than quietly tidied.
Two of the six entries published there are not closed. They say so.
The cost of the rule
The record gets long, and reading it requires knowing that the last verdict is the live one. That is a real cost and it is accepted deliberately: a register that disagrees with reality is worse than no register, because a register is trusted.