Exploratory question
bounded question · negative space or gap · known terms and sources
May not claim: a physical relationship existsEXP01 · EXPERIMENTAL RESEARCH DEVELOPMENT
These improvements connect an exploratory question to preregistration, traceable runs, independent review, replication, bounded claim support, and operational-utility testing. Priority follows dependency and risk.
This register organizes development. It does not authorize a physical experiment, certify a laboratory, approve a protocol, or promote a scientific claim.
Stop automatic progression at the first unmet external, safety, identity, storage, institutional, reviewer, or evidence dependency.
CLAIM MATURITY · E0–E8
bounded question · negative space or gap · known terms and sources
May not claim: a physical relationship existsdeclared variables · detector and frame · ordinary comparison · counterexplanations
May not claim: experimental supportdiscriminating prediction · null world · artifacts · controls · measurement limits
May not claim: authorization or executionfrozen method · analysis plan · exclusions · stop rules · permitted deviations
May not claim: a resultimmutable run identifier · raw evidence references · settings · deviations · failures
May not claim: independent validationmethod review · analysis review · safety record · conflicts · limitations
May not claim: independent replicationfrozen parent protocol · independence statement · comparable evidence envelope · nulls and failures
May not claim: general cross-site supportmultiple independent sites · instrument and frame comparison · conflict preservation · bounded claim language
May not claim: the ontology of Substrate Cosmology is provenreproducible intervention · conventional comparator · benefit and failure bounds · reversibility where applicable
May not claim: fundamental existence beyond the tested scopeEXP02 · PORTABLE CLAIM PACKET
A packet may document approval evidence supplied by named qualified authorities, but this Site contract does not authorize physical execution or certify competence, facilities, ethics, legality, containment, or emergency readiness.
Identifiers and versions are stable; corrections append history rather than replacing prior records.
packet_id · schema_version · version · title · maturity · parent_ids · supersedes · created_on · updated_onA candidate becomes experiment-ready only when outcomes can discriminate among the candidate, ordinary comparison, null world, and measurement artifacts.
question · candidate_relation · conventional_comparison · null_world · counterexplanations · discriminating_predictionsRecord detector substrate, frame, calibration, resolution, environment, operator coupling, and detection limits without treating the detector as neutral.
detector_and_frame · measurement_limits · sites_and_instrumentsFreeze method, variables, controls, exclusions, stopping rules, permitted deviations, and analysis before protected results are known.
protocol · analysis_planRecord hazard class, containment, reversibility, stop authority, permissions, consent, incident path, expiry, and withdrawal; software cannot grant approval.
safetyRaw evidence is immutable; each analysis is a separately versioned branch with code, environment, exclusions, and failure state.
runs · raw_evidence_refs · analyses · uncertainty · limitationsPreserve named review roles, conflicts, independence, failures, and conditions that block comparison.
reviews · replications · conflicts · null_results · non_comparabilityRecord person-specific contributions and AI system, provider, roles, human direction, review, and limitations separately from source authorship or scientific authority.
contributors · ai_assistanceClaim status changes are append-only, evidence-linked, appealable, exportable, and reversible without erasing history.
claim_statement · claim_decisions · corrections · withdrawals · license · export_manifest15 synthetic accepted and rejected cases test missing comparisons, unsupported maturity, unsafe execution authority, mutable evidence, private data, minors, popularity, AI self-approval, and export provenance.
Passing fixtures validates contract behavior only. It does not review a protocol, approve a laboratory, authorize an experiment, validate evidence, or support a scientific claim.
Open synthetic readiness fixtures →EXP01
Public dependency order, maturity ladder, gates, hooks, limitations, and successor paths.
EXP02
Versioned public contract for questions, candidate relations, conventional and null comparisons, protocols, safety, provenance, AI roles, raw evidence, analysis, review, replication, corrections, and claim status.
EXP03
Synthetic accepted and rejected packets that prevent missing controls, unsafe advancement, absent provenance, popularity promotion, AI self-approval, and unsupported claim maturity.
EXP03A
Adult-authenticated public threads can identify E0–E8 maturity, ordinary explanations, important divergence, evidence needed, scalar identifiers, typed replies, reports, and portable public exports.
EXP04
Least-privilege investigator, laboratory, data-steward, method, statistics, safety, ethics, replication, adjudication, and maintenance roles with conflicts and expiry.
EXP05
Owner-scoped drafts, frozen preregistration, immutable run identifiers, auditable changes, export, deletion, retention, and release controls.
EXP06
Hazard-class routing, containment, reversibility, stop authority, institutional permissions, consent, incident reporting, expiry, and withdrawal records.
EXP07
Versioned site, instrument, calibration, resolution, environment, operator-coupling, custody, maintenance, and competence records.
EXP08
Immutable raw-object references, checksums, schemas, code environments, blind analysis branches, uncertainty, exclusions, and reproducible exports.
EXP09
Frozen protocol distribution, declared independence, site matching, evidence envelopes, nulls, failures, deviations, disputes, and non-comparability.
EXP10
Append-only decisions linking exact claim language to protocols, runs, analyses, reviews, replications, conflicts, limitations, corrections, and withdrawals.
EXP11
Reproducible utility comparisons with safety, benefits, failure limits, conventional alternatives, reversibility, and ontology separation.
EXP12
Vendor-neutral schemas, APIs, signed exports, compatibility readers, mirrors, clean-room recovery, migration rehearsals, and retirement paths.
PROTECTED BOUNDARIES
no physical execution authorization · no claim promotion from popularity funding prestige or agreement · no AI system may approve its own scientific contribution · raw evidence is never silently rewritten · nulls failures contradictions and non-comparability remain discoverable · source authorship and AI assistance are separately attributed · student participation remains outside unsupervised experimental workflows · operational utility does not establish ontology