john/ember-memory-test
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

feat(episode): BACKFILL_q-03486d91-69c6-4710-b65d-85ff998743f9_S20260510.BACKFILL-RECOVERED_XX.bkf.0.warm_qdr.orphan.in.cc.es.000.BKF_J.384EE_E.496-3

Browse source 
Skill: NONE | Type: chronicle
Summary: This episode documents the original decision to use UUID v4 as the primary episo
This commit is contained in:
Ember 2026-05-11 04:30:46 +00:00
parent b1f4587529
commit 0c62ec027c
1 changed files with 25 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

25
episodes/BACKFILL_q-03486d91-69c6-4710-b65d-85ff998743f9_S20260510.BACKFILL-RECOVERED_XX.bkf.0.warm_qdr.orphan.in.cc.es.000.BKF_J.384EE_E.496-3.md Normal file
View file

@ -0,0 +1,25 @@
---
episode_id: "384ee496-3a16-55b8-bbaf-a5b016bfd8c3"
puente_flat: "BACKFILL_q-03486d91-69c6-4710-b65d-85ff998743f9_S20260510.BACKFILL-RECOVERED_XX.bkf.0.warm_qdr.orphan.in.cc.es.000.BKF_J.384EE_E.496-3"
session_code: "S20260510.BACKFILL_RECOVERED"
acto_dialogico: "informar"
actor_flujo: "ember_internal"
criticidad_aegis: "moderate"
skill_dominante: "NONE"
fase_proyecto: "maintenance"
tipo_semantico: "chronicle"
summary_one_line: "This episode documents the original decision to use UUID v4 as the primary episode identifier in the EMBER memory system, AND marks that decision as superseded by the PUENTE v2 molecular identifier sy"
source_type: "claude_code"
trust_boundary: "default"
created_at: "2026-05-10T00:00:00+00:00"
relectura_tagged: false
forgejo_commit_sha: "pending"
---
This episode documents the original decision to use UUID v4 as the primary episode identifier in the EMBER memory system, AND marks that decision as superseded by the PUENTE v2 molecular identifier system introduced in session S20260330.
Original decision (S20260329): During the initial MNEMO pipeline design, UUID v4 was chosen as the episode identifier because it provides guaranteed uniqueness without coordination, is natively supported by PostgreSQL (uuid type), Qdrant (point ID), and Neo4j (node property), and is a well-understood standard. The trade-off acknowledged at the time was that UUIDs carry zero semantic information: the identifier 550e8400-e29b-41d4-a716-446655440000 tells nothing about the episode content, type, session, or importance.
Superseding decision (S20260330): PUENTE v2 molecular identifier was designed to address the semantic opacity of UUIDs. PUENTE encodes 21 dimensions across 7 segments (verb, title, session, skills+phase+criticality+decay, dialogue+trust+source+language+flags+scores, AROMA-J, AROMA-E). The PUENTE string is simultaneously an identifier AND a compressed metadata vector that enables filtering, sorting, and partial matching without accessing the full episode payload.
Relectura C1 (contradiction): The original decision stated that semantic identifiers are impractical for production systems because they create coupling between the ID format and the metadata schema. PUENTE v2 contradicts this by demonstrating that a well-designed molecular ID can evolve (v2 to v4.0 added AROMA segments) without breaking existing references, because the parser handles variable-length S5 segments gracefully. Relectura C3 (needs update): The migration plan from UUID to PUENTE documented in the original decision needs updating because the implementation now uses PUENTE as the primary key in PostgreSQL episodes table with UUID retained only as a legacy compatibility field. The ON CONFLICT clause in F5 uses puente_flat (the PUENTE string without underscores) as the deduplication key.