Status: Adopted (agent-recorded design decision) — pending named
architecture ratification Date: 2026-07-24 Source of truth for the data:
evidence/v1-workbenches/studio-authoring-overlap-selection.json (recomputed +
coherence-checked by
node scripts/v1-workbenches/generate-studio-authoring-overlap-selection.mjs)
Grounded in: tara-studio-authoring-overlap.json (§S0.4.a — 315
studio-authoring exports, 9 categories, 0 UI) + tara-import-graph.json (209
tara-workbench exports)
§S0.4.a proved the raw facts; this document records the §S0.4.b–e decisions the inventory cannot make on its own.
S0.4.b — Candidate overlap pairs (exact + near-equivalent)#
- Exact-name duplicates: 0. No symbol is defined in both packages
(
duplicateExportNames: 0). - Concern-level near-equivalence: 117 / 209 Tara exports share a
capability-signal token with a studio-authoring symbol. These are concern
overlaps, not implementation duplicates — the shared tokens name authoring
concepts both packages touch, while the actual symbols differ and Tara's are
meditation-specialized. Top shared concern signals (token → # Tara exports):
source,manifest,gate,block,review,asset,transition,stage,publish,concept,decision,divergence.
S0.4.c/d/e — Selection over the 9 studio-authoring concern categories#
| Category | SA exports | Tara counterpart | Disposition | Basis |
|---|---|---|---|---|
versioning |
22 | src/manifest, src/decision-records |
reuse | version/revision envelope is neutral |
collaboration |
22 | comments/assignments/activity (BFF) | reuse | neutral — matches S0.2 CommentThread/AssignmentPanel |
asset-metadata |
50 | src/bundle, src/manifest |
reuse | accessibility/coverage rollups are neutral |
localization-workflow |
30 | (none) | reuse | fully neutral; Tara has no counterpart (forward reuse) |
creator-roles |
36 | src/review-ladder |
separate | meditation-review roles/grades carry domain trust semantics |
editorial-lifecycle |
55 | src/state-machine, src/gates |
separate | spark→concept→program w/ invitational gates is specialized |
taxonomy-curation |
34 | src/category-registry, src/seasonal-registry |
separate | plane/seasonal taxonomy is meditation-domain |
authoring-blocks |
45 | src/script-blocks, src/invitational-language |
separate | meditation-script authoring (breath/invitational) |
templates |
21 | src/program-schedule, src/seasonal-registry |
separate | cadence/seasonal program templates |
S0.4.c — Direct reuse (4): versioning, collaboration, asset-metadata,
localization-workflow. Each is a domain-neutral concern where Tara should
consume the lower layer rather than re-model. Migration path per row is in the
JSON (.dispositions[].migrationPath) — e.g. reference SA versioning contracts
from the Tara manifest and drop Tara-local version helpers; SA is authoritative
because it is already the lower layer Tara depends on.
S0.4.d — Consolidation (0): none. With 0 exact-name duplicates and a
clean one-way, acyclic layer boundary (Tara → studio-authoring,
studio-authoring does not import Tara — ADR-S1), there is no consolidation
debt: there is no pair of live implementations to collapse into a single
authoritative package. The boundary already prevents the duplication that
consolidation would fix. (This is a real selection outcome — an empty set with a
justification — not a deferral.)
S0.4.e — Retain separate (5): creator-roles, editorial-lifecycle,
taxonomy-curation, authoring-blocks, templates. Each has a Tara
implementation carrying meditation semantics (invitational language, seasonal
registry, program cadence, plane taxonomy, review-ladder grades) that differs
from studio-authoring by domain semantics + audience; both remain valid,
kept separate. The residual Tara-only modules with no SA concern
(src/grounding severity floors, src/steering-overlay, src/provider,
src/errors) are separate by construction.
Consequence#
The two packages are already correctly factored: neutral authoring concerns
belong to @oshun/studio-authoring (the lower layer), meditation specialization
belongs to @oshun/tara-workbench. The forward work is the 4 reuse migrations
(consume SA for neutral concerns), not any de-duplication. This is consistent
with ADR-S1 (staged hybrid) and the S0.2 matrix (extract neutral
interaction contracts; keep Tara vocabulary). Named architecture ratification of
this selection remains a human step; the selection is coherence-gated in CI.
References#
V1_DOMAIN_WORKBENCHES_TODOS_2026-07-23.md§S0.4evidence/v1-workbenches/tara-studio-authoring-overlap.json(§S0.4.a)evidence/v1-workbenches/studio-authoring-overlap-selection.json(this selection, machine-checked)docs/adr/ADR-S1-workbench-kit-package-ownership.md;evidence/v1-workbenches/tara-component-extraction-matrix.md(§S0.2)