DraftReviewed 23 October 2026Reading guide in English

Start with the change.

Current and proposed information

Describe what you observed, what should change and why it matters.

Version and context

Identify the software version, route or stop IDs, language, task and relevant conditions.

Your own checks

Provide a minimal example, expected result and the checks you performed. Disclose substantive AI assistance.

Factual sources and reuse rights

A factual source needs its publisher, exact field or passage, source date and retrieval date. Preserve contradictory evidence and any uncertainty. A file fingerprint identifies bytes; it does not establish truth or permission.

Describe the author, credit, licence and permission for every submitted asset or dataset. Keep credentials, personal data and private customer material out of examples.

Code

Include the package version, a minimal synthetic or open-data input, expected output and the QA report.

Data and editorial corrections

Keep the affected IDs, original claim, precise source and reason for the correction.

Research, language and original assets

Describe the task, meaning, scope, evidence, reuse rights and a way to examine the proposal.

Follow the decision and the output.

These are the proposed research model’s distinctions, not a live public contribution tracker.

Submitted

Received for assessment. This does not imply acceptance.

Accepted

Approved at a specific revision. Publication is a separate step.

Published

Included in an identifiable output or release.

Corrected or withdrawn

Preserve the earlier record and link the replacement, reason and affected output.

Use the technical contribution policy.

Code uses Apache-2.0. Specifications and documentation use CC-BY-4.0. Preserve third-party notices. Sign commits with the Developer Certificate of Origin; a sign-off records your right to submit the contribution and does not replace its source licences.

Choose the available route.

Technical issues are open on GitHub. Other contributions can be prepared locally; the brief workspace saves a file on your device.

Research behind the proposal.

An offline fixture exercise passed 29 checks of selected metadata and lifecycle rules. Its synthetic records are research proof; they are not public submissions, contributor outcomes or production service checks.