EvaluationDetails<boolean> as a downstream receipt boundary #381
Replies: 1 comment
-
|
Small downstream follow-up for context: I landed the Harness-side recipe for this experiment as a runnable path over the existing OpenFeature The boundary is intentionally narrow: boolean Recipe: https://github.com/Rul1an/Assay-Harness/blob/main/docs/OPENFEATURE_DECISION_RECEIPT_PIPELINE.md Keeping this outside the OpenFeature repo; sharing it here only as downstream context for the receipt-boundary idea. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi OpenFeature folks - small downstream build note, no spec change request.
We built an Assay-side compiler lane over a bounded OpenFeature
EvaluationDetails<boolean>artifact: one decision detail row in, one portable evidence receipt bundle out.The interesting bit is the negative space. The receipt is useful because of what it refuses to know. It keeps the small decision surface:
And it deliberately does not import provider config, evaluation context, targeting keys, targeting rules, flag metadata, provider metadata, telemetry, free-text
error_message, or application correctness claims.The larger point: a standard decision surface can become reviewable evidence without turning provider logs, rules, or runtime context into "truth".
This stays outside OpenFeature and is not an official integration or partnership claim. If it is useful context for future docs or discussions around audit/debug consumers of
get*Details, great. Otherwise we will keep it as an Assay-side recipe over the existing decision-detail surface.Links:
Beta Was this translation helpful? Give feedback.
All reactions