Phase 2 — Ingestion and Query Planes (Sketch)

Unit type: Phase (of Epic lattice-platform-development) Unit ID: phase-2 Promotes to: phase-2-plan.md Status record: phase-2-status.md Depends on: phase-1 exit gate

Why this chunk is deliberately light

The epic states its own methodology plainly (Part 0.5, rolling-wave detail): “Phases 2–4 are specified at slice granularity but with lighter test enumeration; each is expanded to full VP-level detail in the last two slices of the preceding phase… Phase 2’s ingestion design will be better informed by what Phase 1 learns about the store SPI.” That is a considered decision, not a gap this decomposition should paper over by inventing detail the epic explicitly says would be premature. Phase 2’s chunk is therefore a real, standalone docs/developer/plans/phase-2.md + status file — satisfying the Epic Decomposition Model’s requirement that every phase gets one — but its slice-level content stays a pointer to epic Part 6 until P1.11.3 (Phase 1’s own close-out slice) expands it.

One piece of Phase 2 is not light: the persistence-compiler integration (P2.1.4a, P2.3.1, P2.3.6, P2.4.1) was already fully designed and written into the epic in a prior decomposition pass, because ontology/persistence/tools/persistence (Slice 2 of rdf-sparql-patterns-phase) already exists and there was nothing premature about specifying how Phase 2 uses it. That work is not repeated here — see epic Part 6’s preamble and docs/developer/INDEX.md Part II §6.

What this chunk still needs to add now, ahead of full expansion

Even a rolling-wave placeholder needs the obligations the user asked every chunk to carry: docs/architecture deliverables, subproject READMEs, and an SDS delta plan for what is already concretely scoped (the persistence-compiler integration, C-05/C-04/C-12’s module boundaries). phase-2-plan.md covers exactly that slice, without inventing VP-level test tables the epic has not yet earned the right to specify.

Risks specific to this phase, worth stating now rather than at P1.11.3

• C-05 (mapping plan compiler) is the largest single piece of missing implementation in the entire epic (Part 14’s critical path calls it out by name) and gates Phases 2 and 3 entirely. Resourcing it first and heaviest is already stated in the epic; this decomposition adds nothing except to note it should not be quietly deprioritised because Phase 2’s chunk looks lighter on paper than Phase 0/1’s.
• P2.9 (content pipeline) explicitly says “Build C-14 first” — the agent orchestrator supervisor — before any LLM call exists in C-06. That ordering constraint belongs in whatever expands this chunk at P1.11.3; flagging it here so it is not lost between documents.