Information Architecture and Retrieval Analysis

Model and evidence

IA designs shared information environments for findability and understanding [EV-BOUNDARY-2026-8C11]. Information-foraging theory explains navigation via cues and expected value [EV-MAINT-2026-7B19]. Developer needs show retrieval must expose rationale and behavior, not merely files [EV-IA-2026-E174].

Counterevidence and interpretations

Findable false information is harmful. Search success does not establish comprehension or authority. More metadata increases maintenance cost. Personalized ranking can hide contradictory evidence.

Reusable methods and outcomes

Findability time, search success, information scent, precision/recall, provenance visibility, stale-result rate, navigation cost, comprehension and action accuracy.

Coverage, gap, and challenged assumptions

IA/IR already explains organization, labeling, retrieval, and cue design. It challenges “put everything in one handoff.” It does not alone validate truth, commitment, or system safety.

Discriminating experiment

Compare self-contained handoffs with indexed evidence-linked handoffs under time pressure and stale-information conditions.

Confidence and limits

High scope overlap; Medium empirical transfer confidence.

Ten future questions

1. Which metadata improves action rather than search only?
2. When is a link better than embedded context?
3. How should stale evidence be ranked?
4. Can contradiction visibility improve calibration?
5. What information-scent cues work across expertise?
6. Does personalization create dangerous omission?
7. Which retrieval failures predict task failure?
8. How should provenance affect ranking?
9. What maintenance cost accompanies richer metadata?
10. Does an integrated profile improve truth/action beyond strong IA?

Completion assessment

Bounded rapid map complete; systematic IR intervention review remains debt.