Point
Resolved processing preserves provenance
A complete processor assembles Point identities, validates cross-record meaning, and emits deterministic records and indexes without flattening their sources.
- Posture
- asserted
- Lifecycle
- active
Resolved processing groups same-named Point files Atlas-wide, requires exactly one anchor, rejects a context in the primary Map, resolves explained memberships to local Areas, checks relation and lifecycle integrity, and derives incoming relations and Map overlap. Every registered Resource URI is resolved exactly once from atlas.md; later Content and Reference uses resolve the Resource id without rereading or re-reporting its target. Resolution therefore depends on the complete discovered Atlas rather than one file at a time.
The normalized model retains the anchor path and every ordered source record with its kind, Map, explained Area memberships, material, extensions, and body. It also preserves every relation note. Collections use deterministic Unicode code-point ordering. Consumers receive assembled identities without losing the semantic edges and contribution boundaries needed to explain the result.
ResolutionThis stage checks anchors, local Area IDs, Resources, relation targets, lifecycle, and cross-record consistency.
Normalized modelThe output preserves explained Area memberships, relation notes, record paths, and derived indexes instead of flattening contributions.