lamina-information-architecture

# Information Architecture (agent-native) Organize **`entities[]`** for actor retrieval — recency, search, hierarchy — not filesystem or schema shape. ## Contract encoding - Entity list drives nav sections and `surfaces[]` groupings - Retrieval patterns per entity type: list, search, timeline, facets — in screen spec - `entities[].relationships` inform parent/child nav — not foreign keys in UI copy ## Frameworks - **Unified object model**: actors manipulate domain objects (tickets, exams), not paths or tables. - **Retrieval**: recency + search + visual browse — match how actors remember work. - **Auto-persist**: no "save?" anxiety — session/history scenarios if undo needed. ## Design checklists 1. IA follows task analysis operations, not org chart. 2. Recent/frequent objects promoted on dashboard screen. 3. Search across entity attributes actors know (name, date) — not internal ids. 4. Deep hierarchy only for rare admin paths. 5. Entity rename in domain → nav label update in same contract pass. ## Verify checks - Actor walk: find entity by primary retrieval path on live product. - "Where did it go?" after create operation → finding if contract promised visibility. ## Anti-patterns - Exposing file paths, table names, module ids to actors. - Folder drilling as only retrieval path. - IA copied from database ERD without task mapping. ## Related - [System Structure](../lamina-system-structure/SKILL.md) - [Navigation](../lamina-navigation/SKILL.md) - [Task Analysis](.....