Alignment Framework
Alignment meta-framework
Section titled “Alignment meta-framework”We align salmon-domain semantics with a top‑down ontology stack so datasets stay interoperable while still retaining domain‑specific meaning.
Hierarchy (top → domain):
BFO → IAO → PROV‑O → SOSA/SSN → I‑ADOPT → Darwin Core → gcdfo:What each layer answers
Section titled “What each layer answers”| Question | Primary ontology | Notes |
|---|---|---|
| What kind of thing is this? | BFO | Material entity, process, role, quality. |
| Is this information or physical? | IAO | Distinguishes information artifacts. |
| Who/what produced the data? | PROV‑O | Provenance, agents, activities. |
| How was it measured? | SOSA/SSN | Observations, sampling, procedures. |
| What property was measured? | I‑ADOPT | Variable decomposition & constraints. |
| How do I publish to GBIF? | Darwin Core | Interoperability scaffold. |
| Domain terms & constraints | gcdfo: | Salmon‑specific classes & labels. |
Variable decomposition (I‑ADOPT)
Section titled “Variable decomposition (I‑ADOPT)”We decompose variables into reusable components:
- Object of Interest — primary entity observed
- Property — characteristic being measured
- Context Object — additional entity providing context
- Matrix — medium/environment
- Constraint — limits on scope (time, size, method)
This helps model variables consistently across datasets and makes mappings precise.
Rules of thumb
Section titled “Rules of thumb”- Check
gcdfo:first. Reuse existing terms whenever possible. - Prefer well‑governed external vocabularies for cross‑domain concepts.
- Never invent IRIs. If a term doesn’t exist, document the gap.
- Use QUDT units for all measures.
Source: salmon-domain-ontology/CONVENTIONS.md