Alignment with CDIF v1.1 Profiles & FAIR Data JSON Schema

Alignment with CDIF v1.1 Profiles & FAIR Data JSON Schema#

This document details the alignment between the FAIR Data JSON Schema dialect and the Cross-Domain Interoperability Framework (CDIF) Version 1.1 specification (published at book.cdif.org).

The FAIR Data JSON Schema and CDIF v1.1 Profiles both pursue the same goal—enabling seamless, cross-domain FAIR data discovery and reuse—approaching the stewardship stack from complementary entry points.

1. What is CDIF Version 1.1?#

The Cross-Domain Interoperability Framework (CDIF), developed by CODATA, WorldFAIR, and the DDI Alliance, defines a set of recommendations and domain-agnostic profiles designed to support cross-domain data integration and reuse.

As documented in CDIF Version 1.1, CDIF establishes standard profiles across key interoperability functions:

  • Discovery Profile: Standardizing core dataset identification (Schema.org / DCAT).

  • Data Description & Structure Profile: Describing tabular and multi-table data structures (DDI-CDI / CSVW).

  • Variable Cascade Profile: Tracing concepts, measurements, and physical columns (DDI-CDI).

  • Provenance & Process Profile: Capturing data lineage, activities, and contributors (PROV-O / DDI-CDI).

  • Access & Licensing Profile: Defining rights, licenses, and access conditions (SPDX / Schema.org).

2. Profile-by-Profile Alignment Matrix#

FAIR Data JSON Schema maps directly to CDIF v1.1 profiles across its Tier 1 (Essential) and Tier 2 (Advanced & Extended) properties.

CDIF v1.1 Profile (book.cdif.org)

Underlying Standards

FAIR Data JSON Schema Mapping

Tier Level

Discovery Profile

Schema.org, DCAT, Dublin Core

$id, $anchor, title, description, fair:label, fair:description, fair:identifiers (identifier, identifierType, uniquenessScope, isResolvable), fair:version (identifier, date), fair:spatialCoverage, fair:temporalCoverage

Tier 1 & Tier 2

Access & Rights Profile

SPDX, Schema.org

fair:license, fair:licenseRef

Tier 1

Contributor & Provenance Profile

PROV-O, DDI-CDI, SKOS

fair:contributors (name, contributorRef, type, role, startDate, endDate), fair:version (notes changelog/rationale -> skos:changeNote)

Tier 1

Data Description & Structure Profile

DDI-CDI, CSVW

fair:resourceType, fair:structureType ("wide", "long", "dimensional", "key-value"), fair:unitType, fair:measurementUnit

Tier 1 & Tier 2

Coded Values & Classifications Profile

SKOS, DDI-CDI

fair:classification / Ref, oneOf + const + title (Hybrid Pattern), fair:sentinel

Tier 1

Variable Cascade Profile

DDI-CDI

fair:conceptualVariableRef, fair:representedVariableRef, fair:instanceVariableRef

Tier 2

Population & Scope Bounds Profile

DDI-CDI

fair:universe / Ref, fair:population / Ref

Tier 2

Data Quality & Measurement Techniques

W3C DQV, ISO 19115

fair:quality (metrics & values), fair:measurementTechnique / Ref

Tier 2

Cross-Dataset Relations & Versioning Profile

DDI-CDI, Dublin Core, OWL

fair:version (identifier -> owl:versionInfo, date -> dcterms:issued), fair:datasetRelations (relationType, targetRef, sourceVariables, targetVariables)

Tier 1 & Tier 2

Catalog Record Provenance Profile

DCAT (dcat:CatalogRecord)

CatalogRecord ($defs/CatalogRecord: conformsTo, sdDatePublished, includedInDataCatalog, about, maintainer)

Refinements

3. How They Work Together (The Ingest Pipeline)#

CDIF v1.1 profiles are traditionally expressed as JSON-LD graphs or SHACL validation shapes. For software engineers and developers, building full JSON-LD / SHACL graphs from scratch can carry a steep learning curve.

FAIR Data JSON Schema provides an intuitive, developer-friendly entry point:

[Developer / Data Pipeline]
         │
         ▼  (Uses familiar JSON Schema Draft 2020-12)
[FAIR Data JSON Schema] ──► Native AI/MCP Integration & REST APIs
         │
         ▼  (Automatic or Steward-guided mapping)
[CDIF v1.1 Profiles (book.cdif.org)] ──► Global FAIR Knowledge Graph & Repositories

The Ingest Workflow#

  1. Developer Ingest: Developers and data owners define their datasets using FAIR Data JSON Schema in standard JSON.

  2. Native AI & API Consumption: AI agents, LLMs, and MCP tools directly read the FAIR JSON Schema for validation and function-calling.

  3. CDIF 1.1 Metadata Generation: Data stewards or automated tools translate the fair: keywords into formal CDIF v1.1 JSON-LD documents for long-term archiving and cross-domain indexing.

4. Technical Comparison Table#

Feature

FAIR Data JSON Schema

CDIF v1.1 Profiles (book.cdif.org)

Primary Format

JSON Schema (Draft 2020-12) with custom fair: annotations.

JSON-LD with SHACL shapes.

Primary Audience

Web developers, API designers, AI engineers, data pipelines.

Institutional data stewards, cross-domain aggregators, archives.

Learning Curve

Zero / Low: Uses familiar JSON Schema syntax.

Medium / High: Requires Linked Data & RDF knowledge.

AI / MCP Readiness

Native: Directly ingestible by LLMs, function-calling, and MCP.

Requires JSON-LD framing or RDF parsers.

CDIF Alignment

100% aligned with CDIF 1.1 Discovery, Structure, and Cascade profiles.

Official CODATA / WorldFAIR specification framework.

5. Summary#

FAIR Data JSON Schema and CDIF Version 1.1 are fully complementary:

  • FAIR Data JSON Schema acts as the developer-facing implementation layer, allowing data creators to capture rich metadata with minimal effort.

  • CDIF v1.1 Profiles act as the global interoperability standard, enabling cross-domain data exchange across global research infrastructures.