Skip to main content

Appendices

This section contains supplementary material to support the Base Product Specification (BPS).
Appendices are informative unless otherwise noted.


A. Extended Meta-Model Diagram

The following extended Mermaid class diagram provides a more detailed view of the BPS meta-model, including optional entities and relationships.


B. Migration Guidance

  • From DPDS/DPROD → BPS:

    • DPDS maps directly to PDS.
    • DPROD maps directly to PROD.
    • Minimal adjustments are required; namespaces should declare conformance to BPS.
  • From Proprietary Schemas → BPS:

    • Proprietary deployment manifests can be mapped into PDS.
    • Proprietary semantic descriptions (e.g., internal catalogs) can be aligned with PROD.
    • Mapping tables SHOULD be documented to ensure interoperability.

C. Optional Profiles

Examples of potential BPS profiles:

  • AIPDS/AIPROD: AI Products.
  • PHPDS/PHPROD: Physical Products.
  • SWPDS/SWPROD: Software Products.
  • Custom Profiles: e.g., STORPDS/STORPROD for storage devices, AUTOPDS/AUTOPROD for automotive products.

Profiles MUST clearly document their scope and relation to the base BPS specification.


D. Extended Examples

D.1 Minimal Example (Data Product)

{
"productId": "urn:example:dataproduct:transactions",
"version": "1.0",
"status": "Published",
"interfaces": {
"input": [{"protocol": "SQL", "format": "ANSI-SQL"}],
"output": [{"protocol": "S3", "format": "Parquet"}]
}
}

D.2 Extended Example (AI Product in JSON-LD)

{
"@context": {
"dct": "http://purl.org/dc/terms/",
"bps": "http://example.org/bps/"
},
"@id": "urn:example:aiproduct:credit-model",
"@type": "bps:Product",
"dct:title": "Credit Scoring Model",
"dct:description": "Predicts applicant creditworthiness.",
"bps:input": "Applicant features (JSON)",
"bps:output": "Credit score (0–1)",
"bps:riskClass": "HighRisk",
"bps:policy": "EU AI Act compliance",
"bps:evaluation": "Bias tested, 2025-01-01"
}

E. Informative Note

Appendices are provided to:

  • Support implementers with diagrams and examples.
  • Help migrators align existing specs with BPS.
  • Show the flexibility of BPS for domain-specific profiles. They are non-normative unless explicitly stated.