Point
Points have Atlas-wide identity
The same Point identifier across Maps denotes one coherent item of context rather than several Map-local duplicates or an entire topic.
- Posture
- asserted
- Lifecycle
- active
An identity names a coherent claim, decision, question, or other item of context, not everything about one subject. A selected storage technology and completion of its rollout need separate Points because their standing can differ. The same decision's operational significance can remain local context.
Independent identity should be useful for reference, relation, or update. A shared posture does not establish identity. Several clauses can explain one assertion whose meaning changes as a unit. Supporting rationale and examples can remain prose. Apply the independent-change test to retained claims that can become false, be implemented, withdrawn, or superseded separately, not to every explanatory sentence. Relevant scope belongs in prose; a clearer explanation need not change the id.
Exactly one record anchors the canonical identity. Other Maps contribute only where local meaning is useful. Exact ids control assembly, while similarity suggests candidates. Validation enforces declared grouping, not the correctness of the author's semantic decision.
Point identity and recordsSame-ID files across Maps resolve together, so identity is Atlas-wide while contribution provenance remains Map-local.
Atlas-wide identity requires authors to search beyond the current Map and choose between a new anchor and a same-named context contribution.
Identity changes the authoring sequence: selecting a Map is not permission to create a Map-local duplicate. The author searches the whole Atlas, creates an anchor only for a new durable identity, and otherwise contributes local meaning under the existing id and filename.
Identity searchAtlas-wide search prevents authors from minting a duplicate identity merely because they started in another Map.
Contribution boundariesOnce an identity is found, the author chooses between changing its anchor and adding a constrained local record.
Processor agreement depends on deterministic grouping of same-named records, one anchor, one record per Map, and retained source provenance.
Atlas-wide identity is observable processing behavior, not only a conceptual statement. Independent processors must assemble the same records into one Point, reject incomplete or conflicting groups, and emit the same primary Map, paths, records, relations, and derived indexes.
ResolutionGrouping by Atlas-wide ID, requiring one anchor, and limiting one record per Map are cross-file resolution rules.
Normalized modelRetained paths and record order make the assembled identity reproducible and auditable.
Both included processors assemble one identity before indexing Maps, relations, and overlap, and their exact normalized outputs must agree.
The reference validator and independent reader discover physical records separately, then group them by Point id and enforce one anchor. Downstream Map indexes and overlap use the assembled identity, which prevents a context file from becoming a duplicate search or navigation entity.
Reference validatorThe reference validator assembles identities before deriving relations and Map indexes.
Independent processingThe second reader must perform the same assembly independently for exact outputs to agree.