Detection Versus Understand¶
Status: Current dimension stage boundary
Authority: Current contract (2026-08-12)
Last reviewed: 2026-08-14
Question¶
Where does dimension Detection end and Understand begin?
Short answer¶
Detection produces mechanical evidence and possible-dimension bundles. Understand assigns engineering meaning and ownership. The Detection UI must not describe possible dimensions as confirmed dimensions.
Detection may claim¶
- text readings that can contain measurements
- arrowhead evidence
- straight-line carriers
- witness-line clues
- feature-contact clues
- possible-dimension bundles with complete, ambiguous, or partial clue sets
0 confirmed dimensionsuntil Understand finishes membership and ownership
Detection must not claim¶
- dimension type as finished meaning
- parsed engineering value and units as accepted truth
- view ownership
- measured-feature ownership
- confirmed dimensions
Understand must assign¶
- dimension type
- parsed value and units
- view ownership
- measured-feature ownership
- ambiguity and confidence
Why the boundary exists¶
The original dimension assembly path mixed association with early meaning. That path became complex and slow. The current contract keeps Detection mechanical and Understand semantic. Incomplete evidence stays visible.
Open Questions¶
- Which ambiguity codes become the stable Understand contract for incomplete dimensions?
Sources¶
- Current contract:
table-dimension-detection-architecture.md(2026-08-12) - Fast Dimension Lines
- Understand