A.1 Model Coherence

James C. Davis

A.1 Model Coherence

A.1.1 The Capability

Keep explicit engineering models connected to both their consumers and the implementation they describe. A model remains useful only when people and agents read the current representation and the implementation continues to correspond to it.

A.1.2 When This Stack Earns Its Keep

Reach for it when:

Engineers or agents rely on explicit system models to reason about the substrate.
Several consumers need the same engineering fact, making copied facts sources of drift.
Implementation can diverge from authored intent with nothing to catch the split.
Downstream artifacts can be generated rather than hand-maintained in parallel.

A.1.3 The Composition

Figure A.1-1. The model-coherence composition. An authored MODEL feeds two paths: consumers CONSUME it live rather than copying its facts, and where the model owns the fact it can EMIT downstream artifacts. Consumption flows into a CORRESPONDENCE check — model against world — which DERIVEs its verdict from stable identities where an independent join exists; on disagreement a GATE gives the modeled property authority. Solid path: the load-bearing composition. Dashed attachment: a useful enhancement, not required for the capability.

A.1.4 Constituent Moves

Table A.1-1.
Move	Role
MODEL	Put the engineering fact in an explicit representation.
CONSUME	Read the authoritative representation instead of copying its facts.
CORRESPOND	Compare model and implementation wherever an independent join exists.
DERIVE	Compute correspondence from stable identities where possible.
EMIT	Generate downstream artifacts when the model can own the fact.
GATE	Reject disagreement where the modeled property deserves authority.

A.1.5 Why These Travel Together

A model nobody consumes is documentation. A model whose consumers copy its facts becomes another source of drift. A model consumed but never checked against reality can go confidently wrong—trusted precisely while it lies.

Direct consumption avoids copied sources of truth; correspondence checks expose divergence between model and implementation; generation eliminates duplicate authority where the model can own a fact outright; and gates make consequential disagreements blocking. Not every model needs every mechanism: a descriptive model may remain advisory. Add stronger mechanisms only when the representation requires stronger authority.

Mechanisms: executable source-of-truth model · meta-model consumption · drift/parity gate · derived traceability · model-derived generation