A node is one rule, and a section holds rules
One numbered rule is one node, and its continuation paragraphs, tables and code blocks belong to that node rather than standing beside it. A section is a container of rules: its text states what the section is responsible for and requires nothing, since every requirement is a rule, and it holds at least one rule — a data record node or a term node counts as one, and the rules of the sections inside it count as its own. A section whose rules have all gone is deleted, and every link and citation naming it is moved to the node that now holds its matter, or removed.
A rule's parts only make sense together. Splitting them into nodes of their own would invent a second address space for fragments nothing can cite. A requirement written into a section's text is a rule nobody can cite, tag or check on its own, and a section with no rule states a responsibility nothing in the code answers to. A data record node and a term node each state on their own what must hold, as a rule does — a data record the shape of its data, a term what its word means in every rule that uses it — and each is cited and tagged by its sid, so a section holding one holds a requirement.
The links made to it
The decisions taken on it
- Problem
The check of a checked section counted a data record node and a term node as its rules, while this rule named rules alone, so a section of terms, such as a glossary of contracts, passed the check and broke the rule as written.
- Ruling
The owner: a data record node and a term node count as rules of their section; that terms are not duplicated is held by contracts alone holding them and by no two of them sharing a name.