Open Data and Crosswalks

Open Data and Crosswalks

Open data workflows should be harvest-first and rights-aware.

Preferred Flow

1. Use a documented protocol or API.
2. Harvest metadata separately from content.
3. Preserve the raw source record.
4. Normalize only after recording identifiers, rights, provenance, and review state.
5. Keep AI interpretation separate from descriptive source metadata.

Crosswalk Fixtures

The repository includes synthetic examples:

• Dublin Core to MARC 21.
• Dublin Core to schema.org.
• Synthetic ledger to synthetic tax worksheet.

Every mapping should declare source path, target path, transform, lossiness, human review requirement, and notes.

Crosswalk source: https://github.com/Protocol-Wealth/pwcli-core/tree/main/crosswalks