Przejdź do treści

Learning Outcome v1

Source schema: doc/schemas/learning-outcome.v1.schema.json

Machine-readable schema for durable correction outcomes produced inside a question-bound answer-room flow.

Governing Basis

Project Lineage

Requirements

Stories

Fields

Field Required Shape Description
schema/v yes const: 1 Schema version.
learning-outcome/id yes string Stable identifier of the correction or confirmation outcome.
question/id yes string Question lifecycle identifier that roots the learning event.
room/id yes string Answer-room or tightly linked review path where the outcome emerged.
subject/ref yes string Reference to the disputed or reviewed answer artifact.
outcome/status yes enum: confirmed, corrected, unresolved Epistemic result of the correction path.
decided-at yes string Timestamp at which the outcome was frozen.
decider/ref yes string Node, secretary, or policy actor that froze the outcome.
supporting/refs yes array Evidence, summary, transcript, or response references that support the outcome.
provenance/refs yes array Trace references sufficient to rebuild the causal chain of the outcome.
summary/ref no string Accepted room summary reference when the outcome is expressed through a summary artifact.
response-envelope/ref no string Corrected or accepted response-envelope reference when a final answer artifact exists.
reason/codes no array Short machine-readable reason tags used by local policy and later audit.
human-linked/input yes boolean Whether the accepted or unresolved outcome depended on human-originated input.
policy/profile no string Optional local or federation policy profile that governed the correction path.
policy_annotations no object Optional implementation-local annotations that do not change the core outcome semantics.

Conditional Rules

Rule 1

When:

{
  "properties": {
    "outcome/status": {
      "const": "corrected"
    }
  },
  "required": [
    "outcome/status"
  ]
}

Then:

{
  "anyOf": [
    {
      "required": [
        "summary/ref"
      ]
    },
    {
      "required": [
        "response-envelope/ref"
      ]
    }
  ]
}

Field Semantics

schema/v

  • Required: yes
  • Shape: const: 1

Schema version.

learning-outcome/id

  • Required: yes
  • Shape: string

Stable identifier of the correction or confirmation outcome.

question/id

  • Required: yes
  • Shape: string

Question lifecycle identifier that roots the learning event.

room/id

  • Required: yes
  • Shape: string

Answer-room or tightly linked review path where the outcome emerged.

subject/ref

  • Required: yes
  • Shape: string

Reference to the disputed or reviewed answer artifact.

outcome/status

  • Required: yes
  • Shape: enum: confirmed, corrected, unresolved

Epistemic result of the correction path.

decided-at

  • Required: yes
  • Shape: string

Timestamp at which the outcome was frozen.

decider/ref

  • Required: yes
  • Shape: string

Node, secretary, or policy actor that froze the outcome.

supporting/refs

  • Required: yes
  • Shape: array

Evidence, summary, transcript, or response references that support the outcome.

provenance/refs

  • Required: yes
  • Shape: array

Trace references sufficient to rebuild the causal chain of the outcome.

summary/ref

  • Required: no
  • Shape: string

Accepted room summary reference when the outcome is expressed through a summary artifact.

response-envelope/ref

  • Required: no
  • Shape: string

Corrected or accepted response-envelope reference when a final answer artifact exists.

reason/codes

  • Required: no
  • Shape: array

Short machine-readable reason tags used by local policy and later audit.

human-linked/input

  • Required: yes
  • Shape: boolean

Whether the accepted or unresolved outcome depended on human-originated input.

policy/profile

  • Required: no
  • Shape: string

Optional local or federation policy profile that governed the correction path.

policy_annotations

  • Required: no
  • Shape: object

Optional implementation-local annotations that do not change the core outcome semantics.