Agenda: content/drafts/dt-overview/trust-chapter.tex¶
Review aid, not a gate. This report is evidence for a human judgement, never a verdict. A driver may read it back; no draft is blocked by what it says. See SOUL.md, and docs/ARCHITECTURE.md's "Layer 4: the review layer".
- Draft:
content/drafts/dt-overview/trust-chapter.tex - Command:
python -m chitragupta.review agenda content/drafts/dt-overview/trust-chapter.tex - chitragupta 6.61.7
How to read this¶
This is a review aid, not a gate: every item below is evidence
for a human judgement, ranked by class and then severity, never a
verdict. unattended items are ones a future agenda-reviser may
act on without asking first; every other item is surfaced for a
person to decide.
Sources¶
- Citation provenance: read
- Verbatim scan: read
- Citation coverage: read
- Multi-source synthesis: read, no item class defined
- TikZ layout check: not run
- Uncited prose: read
- Quotation integrity: read
- Claim support: read
- Prose (style_check): read
- Dossier drift: read
Summary¶
- 2 unsupported-claim
- 16 uncited-claim
Findings¶
unsupported-claim¶
d61be667af63[surfaced] (The argument):[@sample_dt_sync_2023]scores weak: In comparative trials, explicit staleness marking -- returning every value with972e8b4ccf59[surfaced] (Synthesis):[@sample_dt_overview_2024]scores weak: The vocabulary of [...] sharpens the claim: what distinguishes a twin from a sha
uncited-claim¶
d4740a26c3ff[surfaced] (Operational Trust as the Limiting Resource of Digital-Twin Deployments): \label{ch:trust}79e394126f5e[surfaced] (Research question): \label{sec:trust-rq}af107d98e386[surfaced] (Research question): This chapter addresses RQ2: \emph{what limits the operational authority a digitaf5e4c97f0de3[surfaced] (The argument): \label{sec:trust-argument}60febebc8dca[surfaced] (The argument): Three strands of evidence converge on the same answer: the limiting resource is4ff94cf6ac24[surfaced] (Synthesis): \label{sec:trust-synthesis}36ae8634c962[surfaced] (Limitations): \label{sec:trust-limitations}04afcf46eb82[surfaced] (Limitations): The corpus is small and synthetic, its trust measures are behavioural proxies (m7feb3cb1fafd[surfaced] (Limitations): The claim this chapter defends is therefore directional: where the two have been82989c627ad8[surfaced] (The argument): First, the synchronisation literature separates the two directly.03a10e02f0d6[surfaced] (The argument): Accuracy was held constant; honesty moved; trust followed the honesty.9b3b3a830cf2[surfaced] (The argument): Second, the longest observational record in the corpus shows authority followingc607c0693426[surfaced] (The argument): Third, the same asymmetry appears at the analytics layer.902e6461ea7c[surfaced] (The argument): The recommended layered configuration buys accuracy where it is cheap but anchor2b4fb22c185b[surfaced] (Synthesis): Across all three strands, the mechanism that converts a working twin into a trus47496afd7b7d[surfaced] (Synthesis): Authority, in other words, is granted against the override log, not against the