DDS Validation Checklist

This checklist validates dataset readiness under DDS. SFDS suite conformance for DDS is tracked by DDS.manifest.toml and the DDS suite map.

• Dataset manifest exists.
• Sources are recorded.
• License status is recorded.
• Validation procedure is documented.
• Splits are documented when applicable.
• Hashes are recorded for preserved artifacts.
• Known limitations are documented.
• Collection method and transformation steps are documented.
• Removed or filtered records are explained.
• Split method, seed, and counts are recorded when applicable.
• License uncertainty is resolved before public release.
• Dataset class is identified.
• Derived datasets link to source datasets or provenance records.
• Data dictionary or schema notes exist when consumers depend on fields.
• Generated data is labeled and generation method is recorded.
• Sensitive, restricted, or licensed data handling constraints are documented.
• Evaluation sets are protected from accidental training use when applicable.