Dashboard / Workbench Layer / Capabilities / Library

The Library — Capability Architecture

Knowledge organization and retrieval · Work in progress
Status: Active development — two subsystems in progress
Subsystems: Asset Manager (Phase 1 complete), Librarian (4 skills working)
Code: ~4,300 lines Python across 15 scripts
Front of House
Back of House
Collections
Metadata
The Library
Front of House — User-Facing Services
Wayfinding
Orient within the collection. What is here? How is it organized? Where should I look?
Reference Diagnostic
Get behind the question to the actual need. 1–2 clarifying questions, informed by the Mirror.
Facilitate Access
Retrieve, assemble, and contextualize materials. Package for users and other capabilities.
orientation → diagnosis → delivery →
Front / Back Boundary
Back of House — Collection Management
Ingestion
Receive, analyze, describe, and enrich new materials entering the system.
Metadata Manager
Rich metadata scheme for every artifact in the Workbench.
Descriptive Structural Administrative Provenance Schema Evolution
Organizer Live Console →
Build and maintain the organizing systems that make the collection coherent.
Resource Description Classification Relationship Mgt Vocabulary Control Scheme Evolution
Curation
Lifecycle policy — what to promote, archive, discard. Quality thresholds. Collection development.
Collection Management
Integrity, versioning, access control, storage optimization. Executes curation policy.
Collections — By Temporal Character & Use
Reference
Stable, authoritative
Periodical
Current, flowing
Working Files
Active, in-progress
Digital Assets
Media, data
Archives
Completed, preserved
Cross-cutting interfaces: Mirror (user identity, expertise, preferences) · Agency Dial (scaffolded vs. automated retrieval) · Other capabilities (produce artifacts → Ingestion; query → Facilitate Access)
Asset Manager Phase 2 →
Catalog TUG's visual IP — rendering pipeline, extraction, perceptual hash dedup, AI tagging. ~1,690 lines Python, 3 core scripts.
Librarian Skills 4 Skills →
Corpus analysis pipeline: Ingestion → Enrichment → Similarity → Briefing. Grounded in Glushko and Ranganathan.
Organizer Console Live →
Interactive console — Collection Profile, Classification Browser, and Relationship Map for the TUG Blog corpus.