Provenance
Where data came from, and how it was transformed.
Provenance
Provenance in Sounio is not just a logging concern; it is part of how the language wants knowledge-bearing values to remain auditable. The current artifact does not expose the full research model everywhere, but it already enforces an important discipline around intentional boundary crossing.
Current contract
- The clearest provenance boundary in the public surface is still explicit unwrap with a reason.
- Public docs should describe provenance as part of a value’s meaning, not as an after-the-fact annotation.
- When uncertainty and confidence matter, provenance is part of whether a program should be trusted at all.
Where this lives in the repo
docs/reference/KNOWLEDGE_REFERENCE.mdexplains the larger epistemic data model.docs/research/vancomycin-uncertainty.mdshows how provenance and confidence affect a real clinical decision path.self-hosted/check/epistemic.siois the implementation-facing place to inspect current checker logic.
Boundary discipline
let measured = Knowledge { value: 14.2 }
let plain: f64 = measured.unwrap("final report boundary")
Documentation guidance
- When you document a plain numeric result derived from a
Knowledgevalue, explain where and why the unwrap happened. - If a workflow needs stronger provenance guarantees than the checked artifact currently exposes, say so directly.
- Do not collapse provenance, uncertainty, and confidence into one generic “metadata” bucket in user-facing docs.