Everything that has happened to this framework, newest first. Nothing here is ever removed; corrections are new entries.
Conflict notedAIadvisory STD-07 requires every rationale to state checkable sources; these elements carry none: CON-02, STD-04, STD-07, PRF-01, PRF-02
DSC-01Ratifiedhuman + AI Approved by Project steward
DSC-01Proposedhuman + AI Bounded judgment where rules run out, made visible via discretion.exercised / discretion.flagged events.
PRF-02Ratifiedhuman + AI Approved by Project steward
PRF-02Proposedhuman + AI Tiebreaker aligned with INT-01.
PRF-01Ratifiedhuman + AI Approved by Project steward
PRF-01Proposedhuman + AI Tradeable, weighted want — yields when a Standard or Objective demands otherwise.
STD-07Ratifiedhuman + AI Approved by Project steward
STD-07Proposedhuman + AI Dogfood session 2 caught a confabulated rationale via this rule. A rationale trail poisoned with plausible fictions is the one failure mode the system cannot survive.
STD-04Ratifiedhuman + AI Approved by Project steward
STD-04Proposedhuman + AI Rebuild speed is the practical ceiling on schema evolution; slow rebuilds ossify the event catalog.
STD-02Ratifiedhuman + AI Approved by Project steward
STD-02Proposedhuman + AI A universality product that excludes users refutes its own intent.
STD-01Ratifiedhuman + AI Approved by Project steward
STD-01Proposedhuman + AI The event catalog is the contract; schemaless events are unpayable debt in an append-only store.
CON-02Ratifiedhuman + AI Approved by Project steward
CON-02Proposedhuman + AI Honest resourcing bounds the feasible space; violating this constraint is how side projects die.
CON-01Ratifiedhuman + AI Approved by Project steward
CON-01Proposedhuman + AI Boring, proven, operable by a tiny team. XTDB rejected in ADR-0003 on operational-maturity grounds, not capability.
INV-06Ratifiedhuman + AI Approved by Project steward
INV-06Proposedhuman + AI Founding-era caution, explicitly marked as such. Relaxation requires evidence, proposed through the framework itself — not vibes.
INV-05Ratifiedhuman + AI Approved by Project steward
INV-05Proposedhuman + AI Artifact-level attribution failed in dogfood session 2 (the braid problem). Decision roles stay distinguishable under any mixed authorship. Regulators ask who decided, not who typed.
INV-03Ratifiedhuman + AI Approved by Project steward
INV-03Proposedhuman + AI No successful product for non-engineers has ever exposed git. The moment jargon leaks, OBJ-04 is dead.
INV-02Ratifiedhuman + AI Approved by Project steward
INV-02Proposedhuman + AI Authority answers 'says who'; rationale answers 'says why'. Separating and mandating both is the core audit primitive.
INV-01Ratifiedhuman + AI Approved by Project steward
INV-01Proposedhuman + AI If engineers habitually treat history as mutable anywhere, the habit leaks. Redaction events (event.redacted) exist precisely so this never needs an exception.
OBJ-04Ratifiedhuman + AI Approved by Project steward
OBJ-04Proposedhuman + AI The 'git for normies' criterion, stated positively: capability without jargon.
OBJ-03Ratifiedhuman + AI Approved by Project steward
OBJ-03Proposedhuman + AI Extensibility-by-configuration is the falsifiable test of the meta-ontology claim.
OBJ-02Ratifiedhuman + AI Approved by Project steward
OBJ-02Proposedhuman + AI Audit defensibility IS bitemporal answerability: 'what did the framework require' vs 'what did we believe it required'.
OBJ-01Ratifiedhuman + AI Approved by Project steward
OBJ-01Proposedhuman + AI Cross-domain recognizability is the founding UI thesis: position and color carry meaning, vocabulary is skin.
INT-01Ratifiedhuman + AI Approved by Project steward
INT-01Proposedhuman + AI Version control solved trustworthy change for software and stopped there. The same primitives, behind a domain-neutral UI (the ladder), generalize to all normative work. This is the interpretive key for every element below.
Authority registeredhuman + AI Claude Code (ai)
Authority registeredhuman + AI Project steward (human)
Framework createdhuman + AI Ladder (self-governance)