2.8 KiB
name, description, metadata
| name | description | metadata | ||||||
|---|---|---|---|---|---|---|---|---|
| sidecar-typology-structural-floor-vs-protocol-dependent-reading-indexes | The studium engine's two sidecar layers; the .meta.json structural sidecar is protocol-neutral and is the general standard now; reading-indexes are protocol-dependent (chavruta / debate / connection-surfacing) and possibly plural — deferred until the corpus is gold and the protocols declare themselves. |
|
Steward clarification, 2026-06-29 (during Taylor graduation). Two distinct sidecar layers — keep them separate:
1. The .meta.json structural sidecar (studium/meta@1) — PROTOCOL-NEUTRAL, the general standard now.
Says only: where citable text lives, body vs apparatus, sha binding, sections + line anchors. Correct regardless of how the engine is used. This is the bar for "graduated properly" (matches Musil/Position-I precedent). Authored on graduation, sidecars-as-we-go. Lives in studium-engine/corpus/sidecars/<voice-work>.meta.json.
2. The reading-index (rich YAML: core_claims / key_images / key_references / silence_zones per chapter, line-anchored) — PROTOCOL-DEPENDENT, deferred. The Harrison / After-the-Reply YAML indexes were built for the chavruta protocol — a voice that needs to know where to open one book for a given question. The steward's load-bearing doubt: a single species of reading-index probably cannot cover all engine use-cases. Different protocols likely need different interpretive aids:
- chavruta (dialogic, two voices) → where-to-open-the-book map (structure + claims + silence-zones). This is what the existing YAML indexes are.
- full large-scale debate (many voices) → likely a stance / points-of-contention / where-a-voice-pushes-back map, not a where-to-open map.
- connection-surfacing across the corpus → likely cross-text join keys (shared images/references/resonances that bridge works) — closer to MemPalace tunnels than to a per-book ToC.
Discipline: do NOT design a reading-index schema now — premature closure (would encode a chavruta-shaped aid as if universal). Sequence: structural .meta.json for everything → settle the corpus to gold-standard → gather all Making source material → let the protocols declare themselves → THEN design the reading-index species each protocol actually needs. Reevaluate the whole sidecar standard once the corpus is verified gold (steward, 2026-06-29).
Companion thinking: studium-engine/docs/the-sidecar-typology-and-orchestration-2026-06-26.md (PENDING-42 open voice-layer work). The voice-manifest (§VI of the Chamber spec) is a third, still-RESERVED sidecar — distinct from both of the above. See project-making-sequence-source-set, project-studium-engine-telos-chamber-of-voices.