Reader rule
Do not borrow the target geometry
A source-side derivation cannot quietly use the target topology, target atlas, target metric, proper time, detector semantics, or benchmark behavior it is supposed to recover.
Physics deep dive / PG-011
No-Target-Import Discipline
No-target-import discipline prevents a source-side derivation from quietly borrowing the target topology, target atlas, target metric, benchmark success, or process authority it is supposed to justify. It is a guardrail, not completed physics proof.
Reading rule
The guardrail protects derivation integrity.
The discipline is simple: do not use the target structure as a hidden premise, and do not use project-control metadata as mathematical evidence.
Audit rule
The target context is a diagnostic handle
The formal criterion writes C_src(E; T) only to test whether target context changes the construction. T is not an allowed proof input.
Process rule
Receipts are not proof
Validator PASS, registry rows, role identity, handoffs, generated derivatives, local caches, file order, and commits can document process; they cannot prove physics.
Scope rule
The guardrail is not completed proof
No-target-import discipline is necessary methodology. It does not show that all future work is target-import-free or that any candidate is adopted.
Status metadata
What source state this route reflects.
The page explains a control surface and a formal criterion. The current criterion status is `criterion_formalized_pending_audit`.
Forbidden imports
What cannot enter as source evidence.
These categories are blocked because each one can make a source-side proof look stronger than it is by importing target structure or process authority.
Target topology
Open sets and neighborhoods
Target open sets, target neighborhoods, or target manifold topology cannot enter as source premises.
Target atlas
Charts and differentiability
Target coordinate charts, smooth atlases, differentiability classes, and target transition maps cannot be smuggled into source tokens.
Target metric
Metric, signature, and proper time
Metric tensors, Lorentzian signature, causal cones, line elements, distances, and proper-time semantics cannot be used as proof inputs.
Observer semantics
Detector behavior is not source data
Detector response, clock behavior, and empirical observer protocol cannot be treated as source-side proof.
Benchmark success
Exact-GR behavior is not source evidence
Benchmark compatibility can be a later test. It cannot be inserted as a premise for a source construction.
Process authority
Metadata is not mathematics
Validators, registries, roles, handoffs, generated surfaces, local cache state, file order, commits, and recency do not prove a theorem.
Source-side construction
The acceptable path is source-only and fail-closed.
The discipline does not forbid construction. It requires construction to stay inside licensed source data and to stop cleanly when a forbidden dependency appears.
Allowed basis
Use explicit source data
A candidate can use source supports, overlap predicates, transition tokens, certificates, and bottom guards when those are explicitly licensed by the source record.
Factorization
Route proof steps through S(E)
Every non-bottom premise and conclusion should be expressible from source data rather than target or process context.
Replacement
Change T without changing the proof
If target context replacement changes the non-bottom construction, the proof has depended on a forbidden import.
Fail-closed
Record an obstruction instead of continuing
A failed branch should return Bottom_src or record OB-MSRC-TARGET-IMPORT, OB-MSRC-PROCESS-AUTHORITY-LAUNDERING, or OB-MSRC-LOCAL-GLOBAL-GAP.
Public comprehension diagram
The diagram shows filtering, not proof.
The static diagram shows source-only factorization and fail-closed blocking. It is not proof that all future work is target-import-free.
Criterion walkthrough
Read C_src(E; T)=F_src(E) as an audit criterion.
The expression does not prove adoption. It states the test: the candidate written with a diagnostic target-context slot must factor through a source-only construction, or each non-bottom proof step must be expressible without target context.
C_src(E; T)
Candidate with an audit slot
The semicolon marks diagnostic target context: topology, atlas, metric, proper time, detector semantics, benchmark success, and process metadata.
F_src(E)
Source-only construction
This side depends on source data alone. If a proof step cannot be written this way, the criterion fails at that step.
Meaning
Invariance under target-context replacement
Changing the target context should not change the non-bottom construction or proof-step content.
Hard boundaries
What the discipline does not authorize.
The guardrail preserves no `M_src` adoption, no g_eff, no matter coupling, no Einstein equations, no benchmark promotion, no downstream GR promotion, and no completed derivation.
Scope
not proof that all future work is clean
The page explains a discipline. It cannot imply all target imports have been eliminated from all future work.
M_src
no M_src adoption
RT-20260614-128 formalized a criterion pending audit. It did not adopt full M_src.
Metric
no g_eff
No-target-import discipline does not define, construct, adopt, or expand g_eff.
Matter
no matter coupling
The guardrail does not derive matter coupling or adopt a coupling law.
Equations
no Einstein equations
Blocking target imports does not derive Einstein equations.
Promotion
no downstream GR promotion
The criterion does not promote benchmark status or completed derivation.
Specialist source layer
The source records separate guardrail, criterion, audit, and adoption.
Specialist readers should inspect the guard map, criterion artifact, completion record, and claim-boundary rows before treating any no-target-import statement as proof.
RT-105
Guard map control surface
The no-target-import guard map was created as project-system control evidence with no physics claim status change.
RT-128
Criterion formalized pending audit
The artifact defines source-only factorization and target-context replacement invariance for the M_src candidate route.
Review
Criterion theorem is not adoption
The parent conflict review fused mathematical and philosophical views while preserving no M_src adoption and no downstream promotion.
Registry
Boundary language remains upstream
Claim-boundary rows forbid target topology, atlas, metric, proper time, detector semantics, benchmark success as proof, and process authority as proof.
Safe and unsafe readings
A guardrail can be necessary without being sufficient.
This route should improve trust by making the discipline visible, not by pretending the discipline has already solved every future case.
Safe summary
A necessary filter, not a finished derivation.
No-target-import discipline blocks hidden target and process inputs by requiring source-only factorization and fail-closed handling. The current criterion is formalized pending audit.
Unsafe summary
The guardrail proves all candidates are clean.
It would be unsafe to infer all future target imports are eliminated, M_src is adopted, benchmark success is source evidence, or validators and registries can prove physics.
Reader path
Use adjacent pages to separate method from promotion.
Primary reading stays inside the website. Provenance links remain available in the source notice for direct source inspection.
Source extension pipeline
Understand where hidden-import audits sit inside proposal, stress, selector, and human-gated review.
Open routeClaim Boundary Explorer
Inspect source-pinned allowed, forbidden, and gate-required language.
Open routeDistance-to-GR dashboard
Read why methodology criteria do not unlock downstream burdens.
Open routeMetric response ladder
Separate M_src, MetricData(E), g_eff, matter coupling, equations, and benchmark promotion.
Open routeGate Chair and human gates
Review why protected adoption cannot be issued by validators or website prose.
Open route