Skip to content

Context: Lexical

Who is this for (2)

Context: Lexical

Terms describing the glossary’s own internal data model — how terms are represented, resolved, and tracked.

Term Surface

DefinitionThe raw text of a domain term as it appears in mission inputs or outputs, before the system determines what it means. For example, the word “agent” is a term surface that could mean different things in different contexts.
ContextLexical
Statuscandidate
Applicable to1.x, 2.x
Related termsTerm Sense, Glossary Scope

Term Sense

DefinitionThe specific meaning of a term surface within a particular glossary scope. Carries the definition, where it came from (provenance), how confident the system is in it, and its status: draft (auto-extracted), active or deprecated (TermSense and SenseStatus in src/glossary/models.py).
ContextLexical
Statuscandidate
Applicable to1.x, 2.x
Related termsTerm Surface, Provenance, Confidence Score

Provenance

DefinitionMetadata recording where a term sense or artifact originated — who created it, when, and how (e.g., entered by the Human-in-Charge during clarification, extracted automatically from a spec, or seeded from a YAML file). The source field records values such as user_clarification, seed_file and the extraction heuristics (metadata_hint, quoted_phrase, acronym, casing_pattern, repeated_noun).
ContextLexical
Statuscandidate
Applicable to1.x, 2.x
Related termsTerm Sense, Human-in-Charge (HiC)

Confidence Score

DefinitionA value between 0.0 and 1.0 indicating how certain the system is about a term extraction or sense resolution. Higher values mean higher certainty. It feeds conflict severity scoring (score_severity in src/glossary/conflict.py): an unknown term with confidence below 0.5 at a critical step scores high severity, which blocks generation under medium strictness.
ContextLexical
Statuscandidate
Applicable to1.x, 2.x
Related termsTerm Sense, Semantic Check