Interoperability Layer, Not Replacement Layer
The Principle
When multiple classification systems exist and can’t be consolidated, create a shared reference vocabulary that each system can map to rather than attempting to replace them.
Why This Matters
In any domain with multiple stakeholders, you’ll find multiple taxonomies. Organizations build classification systems that reflect their local context, regulatory requirements, and historical decisions. These systems have:
- Sunk investment (tooling, training, integrations)
- Embedded local knowledge
- Political constituencies defending them
Mandating a single replacement system creates resistance and loses valuable local context. The alternative: create a reference layer that enables translation between existing systems without requiring abandonment.
How to Apply
- Design the reference taxonomy for breadth and mappability, not perfection
- Create crosswalk tables linking local concepts to reference concepts
- Support multiple match types (exact, broader, narrower, related)
- Maintain bidirectionally: changes in either direction need to propagate
Example (ESCO): Each EU member state maintains national occupational classifications. ESCO provides a common vocabulary. Spain’s system maps to ESCO. Latvia’s system maps to ESCO. Now Spanish and Latvian job platforms can interoperate without either country changing their internal systems.
When This Especially Matters
- Cross-border or cross-organization data exchange
- Federated systems without central authority
- Legacy system integration
- Any domain with regulatory or political fragmentation
Exceptions
If you have clear authority to mandate a single system and local context truly doesn’t matter, a unified taxonomy is simpler. Interoperability layers add maintenance overhead and mapping ambiguity.
Related: [None yet]