List of Figures and Tables

James C. Davis

List of Figures and Tables

The figures and tables of this book, in order. Each links to where it appears. Every figure, rendered in full with its caption, also lives on the Figures Gallery — a one-page visual review of the whole book.

Figures

Figure 0.1-1 — The book's argument, from the Preface's opening question through the seven…
Figure 0.3-1 — The MAGE method
Figure 0.4-1 — The six model classes, organized by the engineering questions they answer
Figure 0.4-2 — The agent stack and a characteristic failure mode
Figure 0.4-3 — The governed environment
Figure 0.5-1 — From case to theory
Figure 1.1-1 — Where engineering leverage acts
Figure 1.2-1 — DocAble from the user's view
Figure 1.2-2 — Processing one document in DocAble
Figure 1.3-1 — Agentic machinery and environment
Figure 1.3-2 — The reasoning horizon
Figure 1.4-1 — From substrate to method
Figure 2.1-1 — Obligations bound a realization space without selecting one realization
Figure 2.1-2 — From engineering question to engineering consequence
Figure 2.1-3 — The engineering modeling repertoire
Figure 2.1-4 — MAGE's working model ontology
Figure 2.2-1 — The structural move
Figure 2.2-2 — Document-mutation structural model
Figure 2.2-3 — From computations to composition
Figure 2.3-1 — The behavioral move
Figure 2.3-2 — Recovery-ordering model
Figure 2.3-3 — Behavioral models expose different classes of properties
Figure 2.4-1 — The ownership move
Figure 2.4-2 — In-flight ownership model
Figure 2.4-3 — Model classes overlap
Figure 2.5-1 — The decision move
Figure 2.5-2 — Service-flow model
Figure 2.5-3 — Observation and intent answer different questions
Figure 2.6-1 — The measurement move
Figure 2.6-2 — Core measurement relation
Figure 2.6-3 — GenAI cost-and-capacity model
Figure 2.6-4 — Modeling and enforcement are separate decisions
Figure 2.7-1 — From observability to provenance
Figure 2.7-2 — Attribution provenance model
Figure 2.7-3 — Computation structure versus realized history
Figure 2.8-1 — Heterogeneous models over shared identities
Figure 2.8-2 — Task-driven traversal
Figure 2.8-3 — Three correspondence patterns
Figure 2.8-4 — Coverage of model correspondence
Figure 2.8-5 — Checked predicates over a structured model
Figure 3.0-1 — How intent becomes enforceable
Figure 3.1-1 — Intervention boundaries
Figure 3.2-1 — Three distinct relations
Figure 3.2-2 — Correspondence strength and governing role are independent
Figure 3.2-3 — The semantic gap
Figure 3.3-1 — Constraint or sensor
Figure 3.3-2 — Match the mechanism to the property
Figure 3.3-3 — One Measurement, Different Enforcement Decisions
Figure 3.3-4 — Negative and positive constraints
Figure 3.3-5 — Provenance-carried admission
Figure 3.4-1 — Governance conversion
Figure 3.4-2 — Correct judgment without enforcement
Figure 3.4-3 — Where enforcement stops
Figure 3.5-1 — The conflict is the edge
Figure 3.5-2 — Three views of the control machinery
Figure 3.5.1-1 — Two execution units for document remediation
Figure 3.5.1-2 — Chunked residency versus skeletonized residency
Figure 3.5.1-3 — Two execution policies over the same remediation graph
Figure 3.5.1-4 — Analysis before implementation
Figure 3.5.1-5 — The GenAI dependency chain and its latency floor
Figure 3.5.1-6 — Using prediction error as evidence
Figure 3.5.1-7 — Modeling as a loop
Figure 4.1-1 — Precision follows knowledge — uncertain need → explore → stabilizing claim → represent → govern where justified
Figure 4.1-2 — Engineering recurring work
Figure 4.2-1 — Where a migration starts
Figure 4.2-2 — Representation and enforcement strengthen independently
Figure 4.2-3 — Discovering models during migration — weak signals → partition {legitimate shape · mechanical debt · latent concept} → name the model → migrate consumers → narrower enforcement → stronger substrate
Figure 4.2-4 — Audit, Drain, Promote
Figure 4.2-5 — Derive, don't copy — one source of truth, consumers query/join/generate; where neither determines the other, parity check
Figure 4.3-1 — A Point and a Space
Figure 4.3-2 — Let the claim choose the search
Figure 4.3-3 — From models to evidence — model/spec → derive obligations → census → choose evidence (test/lint/search-proof) → claim coverage
Figure 4.3-4 — Generate, Judge, Search Again
Figure 4.3-5 — Two evidence boundaries — earliest legible boundary catches cheaply near the cause; last safe boundary re-establishes what must hold
Figure 4.4-1 — Lifecycle event → deterministic fire → execute / delegate / escalate → next state
Figure 4.4-2 — Externalizing judgment — recurring procedure → three kinds (Execute/Delegate/Escalate) → three forms (tool/prepared brief/human)
Figure 4.4-3 — Surrounding a judgment step with deterministic work
Figure 4.5-1 — Domain model → recurring concerns → decision procedure → reusable skill
Figure 4.5-2 — Two layers of context — standing context (always resident) and a task-relevant slice retrieved on demand feed the agent
Figure 4.5-3 — MAGE as agent knowledge
Figure 5.0-1 — Two views of the evidence
Figure 5.2-1 — The Seven Build Stages
Figure 5.2-2 — Growth of Two Countable Control Artifacts
Figure 5.2-3 — The Support Ratio
Figure 5.3-1 — What Got Built
Figure 5.3-2 — Model as a Bounded Subroutine
Figure 5.3-3 — The modeling history
Figure 5.3-4 — One Seam, Hardened in Layers
Figure 5.3-5 — The Delegation Staircase
Figure 5.4-1 — Modeled and Observed, Not Yet Binding
Figure 5.5-1 — Eight Entry Points
Figure 5.5-2 — The Modeling Ceiling
Figure 6.1-1 — The MAGE Dynamic Model
Figure 6.1-2 — The Determinization Frontier
Figure 6.3-1 — The economics of explicit engineering
Figure 7.1-1 — Where Engineering Effort Moves
Figure 7.2-1 — Why software modeled differently
Figure 7.2-2 — Established traditions, new composition
Figure A-1 — Seven reference engineering stacks and their common relationships
Figure A.1-1 — The model-coherence composition
Figure A.2-1 — The assurance composition
Figure A.3-1 — The auditable-transformation composition
Figure A.4-1 — The observe → react composition
Figure A.5-1 — The resource-mediation composition
Figure A.6-1 — The governance-conversion composition, a loop
Figure A.7-1 — The context-delivery composition
Figure A.8-1 — Composing an engineering stack
Figure B-1 — Ten recurring engineering problems and the moves that address them
Figure B.1-1 — One authoritative representation
Figure B.2-1 — Query, don't snapshot
Figure B.3-1 — Correspondence runs both ways
Figure B.4-1 — Required set minus present evidence
Figure B.5-1 — Earliest legible, last safe
Figure B.6-1 — Open surface versus closed seam
Figure B.7-1 — Knowledge delivery at the decision point
Figure B.8-1 — Cause travels with consequence
Figure B.9-1 — From system model to operational guidance
Figure B.10-1 — Impact as a graph query
Figure C-1 — The executable-model pattern
Figure C.1-1 — Structure and boundaries
Figure C.2-1 — Behavior plus ownership
Figure C.3-1 — Stable topology, variable execution policy
Figure C.4-1 — Measurement does not imply enforcement
Figure C.5-1 — Facts before prose
Figure C.6-1 — Composition by reference
Figure D.9-1 — Brownfield progress
Figure E.2-1 — Three complementary skills act around one governed engineering environment
Figure F.1-1 — MAGE across the product lifecycle
Figure F.4-1 — Change-scoped models, degrees of freedom, and inheritance
Figure F.5-1 — From incident repair to governance conversion
Figure F.6-1 — Assurance across models and realization
Figure F.8-1 — From local MAGE adoption to the product GEE
Figure F.9-1 — Engineering capital across time and space
Figure G.1-1 — From assistance to bounded delegation
Figure G.2-1 — From recurring cost to engineering capital
Figure G.2-2 — Four figures this appendix puts to work
Figure H-1 — Cloudflare projected onto MAGE
Figure H-2 — Spotify projected onto MAGE
Figure H-3 — Shopify projected onto MAGE
Figure H-4 — Docker projected onto MAGE
Figure H-5 — Siemens projected onto MAGE
Figure H-6 — Zenseact projected onto MAGE
Figure H-7 — Uber projected onto MAGE
Figure I-1 — Weekly Commit Volume
Figure I-2 — Product-Path Line Motion
Figure J.2-1 — Model induction

Tables

Table 4.2-1 — Sizing the mechanism
Table 4.5-1 — Governance versus operate
Table 5.2-1 — The build's scale at one snapshot
Table 6.1-1 — The three MAGE outcomes and their observables
Table 6.1-2 — The four Modeling × Alignment combinations
Table 6.2-1 — MAGE's nine hypotheses in three families
Table 6.3-1 — Typical MAGE profiles by setting
Table 7.3-1 — The authorship division
Table 9.0-1 — The practice appendices and the questions they answer
Table C.1-1 — Component-zone properties
Table C.2-1 — Lifecycle invariants
Table C.7-1 — Representation to enforcement receipt
Table D.1-1 — The Operator's Dashboard — five primary readings
Table D.8-1 — Joining the wiki to the code
Table D.8-2 — Audit, synchronize, govern, extend
Table 17.0-1 — The evidence appendices and the questions they answer
Table 21.1-1 — The manuscript's governing models and the failures they prevent