Removal goes through the archive when the node was ever confirmed
A node a confirmation was ever given for is moved to the archive when it is removed, together with the reason for its removal and the sid of the node that replaces it, if one does. A node never confirmed is deleted outright, whatever cited it.
A node may later be deleted from the archive.
A rule its owner confirmed is one the code may implement, and a later change to its wording does not undo what was implemented from it, so its removal is worth keeping for a time. Kept in its last wording, it can be restored if its removal is undone. And it is a record for the decisions taken: a confirmed rule that had to go, a rule of the base above all, most often shows a logic or an approach that was mistaken, which later work has to understand and take into account, as it would a list of the defects found.
Nothing guarantees how long the archive keeps a node, so a link or a citation naming an archived node breaks when the node is deleted, and the checks refuse it until it is removed. A node never confirmed was never implemented, since no tag may name it, and archived it would only bury the rules that had to go under wording nobody confirmed.
The links it makes
The links made to it
The decisions taken on it
- Problem
The reason spoke of the specification's owner, while a confirmation is now given by an owner of the node's part.
- Ruling
The owner: an owner of the node's part (§s-67vekyup7y8smap).
- Problem
The reason held that the archive keeps resolvable the citations in the code, the tests and the commits that implemented a removed rule, while nothing keeps a node in the archive for good.
- Ruling
The owner: the archive keeps a removed node for a time, so that it can be restored and so that later work learns from what was found wrong, as from a list of the defects found; a node may be deleted from it, and a link into it may then break.