journal: DOLCHEO TSVs for 2026-05-11 sessions (exploration + drafting)

[klappy]

Two structured DOLCHEO journal files in the established journal/*.tsv format (8 columns: type, typeName, facet, quality_score, quality_max, quality_level, title, content).

Files

• journal/2026-05-11-agent-runtime-exploration.tsv — 9 artifacts from the exploration session:

  • architecture-already-canon observation (5/5 strong)
  • three-layer-architecture constraint (5/5 strong)
  • one-loop-many-surfaces handoff correction (5/5 strong)
  • CF-Agents-Week-alignment constraint (5/5 strong)
  • seven-canon-gaps handoff (5/5 strong)
  • two-dispatch-paths constraint (5/5 strong)
  • Bide-criterion-re-cut decision (5/5 strong)
  • R2/object-store trigger observation (5/5 strong)
  • catalog-and-orient workflow learning (4/4 strong)

• journal/2026-05-11-trigger-taxonomy-drafting.tsv — 10 artifacts from the drafting session:

  • dispatch-routing-layer decision — post-challenge structural revision (5/5 strong)
  • structural plan handoff (4/4 strong)
  • surface-vs-trigger constraint (3/4 adequate)
  • trichotomy constraint + decision pair (4/4 + 5/5 strong)
  • nine-source observation (4/4 strong)
  • triple-redundant-handoff-overstatement learning (4/4 strong)
  • word-count-discipline-failure learning (4/4 strong)
  • mode-collapse-into-process-narration learning (4/4 strong)
  • Move-A-complete handoff (5/5 strong)

Pairing with narrative ledger

Both TSVs are companions to the narrative MD ledgers at odd/ledger/2026-05-11-*.md (PR #199). The MD captures the causal narrative (what led to what); the TSV captures the encoded DOLCHEO artifacts in queryable form. The title column is the join key between formats.

Why both formats

Per the operator's framing: TSV powers awk / grep / queryability across the project's full DOLCHEO history (awk -F'\t' '$1=="D"' journal/*.tsv returns every Decision; quality filtering trivial; aggregation by epoch / persona / quality tier viable). MD preserves the causal chain that TSV-by-row destroys, per odd/ledger/project-journal-best-practices.md ("Separating by type destroys the causal chain"). Both consumers — human readers and tooling — get the format they need.

Schema

Verified against existing files: journal/2026-05-10-runtime-contract-reconciliation-and-governance-override.tsv, journal/2026-05-10-ams-canon-alignment-pass.tsv, journal/2026-05-08-oddie-bootstrap-wiring.tsv. 8-column header, tab-separated, one row per encoded artifact.

---

Note

Low Risk
Low risk: adds new journal/*.tsv artifacts only (no runtime/canon code changes), so impact is limited to documentation/queryability.

Overview
Adds two new journal/*.tsv DOLCHEO entries for 2026-05-11.

The new artifacts record conclusions from an agent-runtime exploration session (layer separation, dispatch-path framing, identified canon gaps, and object-store triggers) and a trigger-taxonomy drafting session (dispatch-routing-layer decision, surface-vs-trigger constraint, routing-method trichotomy, trigger-source enumeration, plus process learnings and a handoff referencing prior substrate-options work).

^{Reviewed by Cursor Bugbot for commit d4c8482. Bugbot is set up for automated code reviews on this repo. Configure here.}