A link may cross a domain line
A link between nodes of the specification may cross a domain line: a rule of one domain may point at a rule of another, and the link carries its type like any other. Whoever would point at a statement of another domain that no node there holds first gives it to a node of that domain that can carry it, or, failing one, writes a rule there whose todos name the node linking to it, and links to that node.
Domains build on one another — a later one uses what an earlier one provides — and the specification has to be able to say so. A link holds the sid of the node it points at, so moving either node leaves it whole, and a link to a node that no longer exists is refused by the checker the same day. A statement named in words rather than by a node is a reference no check can follow, and the node it means shows nobody relying on it; given to a node there it keeps one home, and a rule written for a link is read when its domain is, its todo saying why it stands there.
The links made to it
The decisions taken on it
- Problem
The citation check refused a citation crossing a domain line unless it named a node of a few domains, a list the tool kept and no rule stated, and such citations had been replaced by the words of the rule they named; that a citation into another domain names a node there stood nowhere.
- Ruling
The owner: a link of any type, cites among them, crosses a domain line freely and is checked only for naming a node, over admitting one only into what the linking domain's header consumes; a statement no node of the other domain holds is given to a node there, or a rule is written there for the link.