RFC-0052 platform domain review
Versioned draft snapshot
RFCs, reference catalogs, and evidence reflect Spec 3b2e3e7a114fb826b1445d39a48756b8fff629cd. Counts and pending/status notes belong to that snapshot. Read the dated project update for later progress.
Source repositories are private during staging. Links marked “repository access required” need access to the linked repository.
This is a trace and gap record from the pinned draft specification. It does not establish proof or implementation conformance.
Status: all 74 domain nodes reviewed on 2026-10-05; 59 match and 15 mismatch. Owners: Platform, Compiler, Security. No RDF facts changed.
Thirteen core.* nodes match §4.1’s named bootstrap declarations. They are specific structural meta-gnomon kind definitions, distinct from generic DSL forms and physical PostgreSQL tables/functions.
The private/protected/public template visibility values are typed Capability. Visibility modifiers are not separately grantable capabilities. Introduce a specific visibility value class or link to the canonical visibility type after reviewing its other owners. Protected visibility has an additional source conflict: §8 describes descendant-only access but permits core.* to use protected db.graph templates, although core.* is not a db.graph descendant.
Acceptance: identify the visibility owner and exact exception model, preserve capability checks separately from visibility resolution, reconcile source prose and Lean, update RDF classes with owner approval, and test private, descendant-protected, explicitly allowed core, and public callers against the same resolver. Source review does not establish operational enforcement.
Named template and type declarations
Twenty-four more nodes were reviewed. Nineteen agree with explicit core kind, platform template, db.pg type, and DDL-template catalog entries. The named source templates remain distinct from emitted database objects.
Five column-aspect summaries exceed the catalog evidence: timestamps, soft-delete, audited-creator, data-jsonb, and bigint-pk. Their NamedDeclaration classes match their named aspect identities, but §3.1 lists names without formal contribution schemas. The summaries add exact fields or behavior that need authored definitions or provenance to a defining RFC/module. Class agreement alone does not make the whole node match.
No source facts changed. Acceptance requires concrete aspect contributions, consistent inheritance/merge rules, owner attribution, generated projection updates, and tests of emitted columns when codegen is available.
Remaining catalog nodes
The final 34 nodes contain 27 matches and seven mismatches. The named base-type and DDL-template declarations, module catalog entries, and six remaining invariants agree with the authored catalog.
Five terms (ApplicationModule, BootstrapModule, InstallRank, LayeredModule, and PlatformModule) are catalog concepts, but are projected as DataType. The ontology reserves DataType for primitive or named Gnomon language types; §2 supplies definitions without declaring such types. Acceptance requires a catalog-concept or module-role representation, an inventory of consumers, and owner approval before changing the classes.
The single-schema invariant copies §10.2 but conflicts with §3: db.graph owns graph and db.ids owns ids, rather than db_graph and db_ids. The dot-to-underscore rule belongs to application view schemas in §5. Preserve the explicit platform schema map and db.pg’s schema-less exception when reconciling the narrative, Lean, and RDF.
The bootstrap invariant also exposes an authored inconsistency. §10.4 and Lean claim convergence with Phase 2 on an empty database, while §6 supplies Phase-1 prerequisites and §6.1 describes repeat-install idempotency on a bootstrapped database. Define these as separate properties and specify any fresh-install prerequisite path before claiming convergence. Acceptance requires tests for bootstrap followed by install, repeat install, and the explicitly supported fresh-install path.
These findings record source disagreements; they do not change normative facts or demonstrate operational enforcement.