Context: Practices and Principles
Terms describing operating principles for glossary-semantic-integrity execution.
Mostly Invisible Process
| |
|---|
| Definition | Glossary checks should run by default with minimal interruption during normal flow. |
| Context | Practices & Principles |
| Status | canonical |
| Applicable to | 1.x, 2.x |
Highly Visible Outcomes
| |
|---|
| Definition | Conflict results, block reasons, and clarification decisions must be visible in artifacts/logs. |
| Context | Practices & Principles |
| Status | canonical |
| Applicable to | 1.x, 2.x |
Append-Only Semantic History
| |
|---|
| Definition | Term/sense changes and conflict decisions are recorded as immutable events, not rewritten in place. |
| Context | Practices & Principles |
| Status | canonical |
| Applicable to | 1.x, 2.x |
Deterministic Gate, Human Clarification
| |
|---|
| Definition | Conflict detection/gating is deterministic; only ambiguity resolution requires human input. |
| Context | Practices & Principles |
| Status | canonical |
| Applicable to | 1.x, 2.x |
High-Severity Block Rule
| |
|---|
| Definition | Under the default medium glossary strictness, generation is blocked only when high-severity semantic conflicts remain unresolved. max strictness blocks on any unresolved conflict and off never blocks (should_block in src/glossary/strictness.py). |
| Context | Practices & Principles |
| Status | canonical |
| Applicable to | 1.x, 2.x |
Human-in-Charge (Cross-Reference)
| |
|---|
| Definition | The principle that the human remains responsible and accountable for all decisions. Agents assist and propose; the HiC owns the final call. See full definition in Identity context. |
| Context | Practices & Principles |
| Status | canonical |
| Applicable to | 1.x, 2.x |
| Canonical entry | Human-in-Charge (HiC) |