Direct answer
Reproducibility — Machine Record is implemented as a source-bounded machine record guide.
Answer contract
Apply the machine record lens only to the exact inspected source scope; do not infer authority from adjacent topics.
Evidence and exact locators
t23-reproducibility-registry — lib/evidence-workflow-examples.ts — EVIDENCE_WORKFLOW_PUBLIC_REGISTRY. Supports: Publishes digest-bound synthetic workflows, refusals, and checks.
nasem-reproducibility-2019-t9 — https://nap.nationalacademies.org/catalog/25303/reproducibility-and-replicability-in-science — Consensus Study Report, 2019; Chapter 3, Understanding Reproducibility and Replicability; report summary definitions. Supports: Computational reproducibility means obtaining consistent computational results with the same data, steps, methods, code, and analysis conditions; replicability uses new data to answer the same scientific question.
local-computational-witness-repro-t9 — repo:packages/maha-witness/README.md — repository source inspected 2026-09-06; Recorded fields; Adapter boundary; Assurance boundary; verification command. Supports: The witness records execution identity, bounded environment metadata, seeds, configuration, and artifact hashes, and verifies the receipt digest offline.
What the evidence does not establish
Synthetic conformance is not independent reproduction.
Reproducibility does not establish scientific correctness or external validity, and a failed replication does not by itself identify which study or condition is wrong.
It does not capture arguments, arbitrary environment variables, paths, file contents, or credentials and does not establish scientific correctness, complete environment capture, or independent reproduction.
Rights and reuse
project-owned-reference-only
Publicly accessible source; link and bounded paraphrase only. No source text is redistributed.
Maha-controlled repository source; headings, symbols, and file identity only. No credential, customer value, runtime payload, private input, or receipt body is retained.
Dependencies and related concepts
applies-to: urn:maha:concept:evidence:reproducibility
governed-by: urn:maha:concept:governance
evidence-for: urn:maha:concept:evidence
Source or implementation
This source or implementation section is constrained to the same inspected scope: Publishes digest-bound synthetic workflows, refusals, and checks. Computational reproducibility means obtaining consistent computational results with the same data, steps, methods, code, and analysis conditions; replicability uses new data to answer the same scientific question. The witness records execution identity, bounded environment metadata, seeds, configuration, and artifact hashes, and verifies the receipt digest offline.
It must preserve the recorded boundary: Synthetic conformance is not independent reproduction. Reproducibility does not establish scientific correctness or external validity, and a failed replication does not by itself identify which study or condition is wrong. It does not capture arguments, arbitrary environment variables, paths, file contents, or credentials and does not establish scientific correctness, complete environment capture, or independent reproduction.
Method
This method section is constrained to the same inspected scope: Publishes digest-bound synthetic workflows, refusals, and checks. Computational reproducibility means obtaining consistent computational results with the same data, steps, methods, code, and analysis conditions; replicability uses new data to answer the same scientific question. The witness records execution identity, bounded environment metadata, seeds, configuration, and artifact hashes, and verifies the receipt digest offline.
It must preserve the recorded boundary: Synthetic conformance is not independent reproduction. Reproducibility does not establish scientific correctness or external validity, and a failed replication does not by itself identify which study or condition is wrong. It does not capture arguments, arbitrary environment variables, paths, file contents, or credentials and does not establish scientific correctness, complete environment capture, or independent reproduction.
Limits and uncertainty
This limits and uncertainty section is constrained to the same inspected scope: Publishes digest-bound synthetic workflows, refusals, and checks. Computational reproducibility means obtaining consistent computational results with the same data, steps, methods, code, and analysis conditions; replicability uses new data to answer the same scientific question. The witness records execution identity, bounded environment metadata, seeds, configuration, and artifact hashes, and verifies the receipt digest offline.
It must preserve the recorded boundary: Synthetic conformance is not independent reproduction. Reproducibility does not establish scientific correctness or external validity, and a failed replication does not by itself identify which study or condition is wrong. It does not capture arguments, arbitrary environment variables, paths, file contents, or credentials and does not establish scientific correctness, complete environment capture, or independent reproduction.
Dependencies
This dependencies section is constrained to the same inspected scope: Publishes digest-bound synthetic workflows, refusals, and checks. Computational reproducibility means obtaining consistent computational results with the same data, steps, methods, code, and analysis conditions; replicability uses new data to answer the same scientific question. The witness records execution identity, bounded environment metadata, seeds, configuration, and artifact hashes, and verifies the receipt digest offline.
It must preserve the recorded boundary: Synthetic conformance is not independent reproduction. Reproducibility does not establish scientific correctness or external validity, and a failed replication does not by itself identify which study or condition is wrong. It does not capture arguments, arbitrary environment variables, paths, file contents, or credentials and does not establish scientific correctness, complete environment capture, or independent reproduction.
Questions this page can answer
What is established?
Reproducibility — Machine Record is implemented as a source-bounded machine record guide.
Which exact source establishes it?
t23-reproducibility-registry, lib/evidence-workflow-examples.ts — EVIDENCE_WORKFLOW_PUBLIC_REGISTRY; nasem-reproducibility-2019-t9, https://nap.nationalacademies.org/catalog/25303/reproducibility-and-replicability-in-science — Consensus Study Report, 2019; Chapter 3, Understanding Reproducibility and Replicability; report summary definitions; local-computational-witness-repro-t9, repo:packages/maha-witness/README.md — repository source inspected 2026-09-06; Recorded fields; Adapter boundary; Assurance boundary; verification command
What dependency remains separate?
Synthetic conformance is not independent reproduction. Reproducibility does not establish scientific correctness or external validity, and a failed replication does not by itself identify which study or condition is wrong. It does not capture arguments, arbitrary environment variables, paths, file contents, or credentials and does not establish scientific correctness, complete environment capture, or independent reproduction.
What uncertainty remains?
applies-to: urn:maha:concept:evidence:reproducibility governed-by: urn:maha:concept:governance evidence-for: urn:maha:concept:evidence
What must not be inferred?
A source, locator, rights, scope, boundary, dependency, implementation, or release change requires a new exact-revision review.