Delivery Timeline
| Milestone | Effort (BE/FE/QA…) | Confidence | Progress | Target Date | Status |
|---|---|---|---|---|---|
| PRD finalized | — | — | 0/0 | TBD | Pending — PRD (historical-validation.md) is still draft |
| RFC kickoff | — | — | 1/1 | 2026-06-20 | Done — RFC drafted and reviewed three cycles (historical-validation-review.md R1→R3); R3 is a re-grounding refresh with zero citation drift, score holds 8.5/10, verdict PROCEED (only 5 minor open findings: REV-5/6/7/8/9, no blockers/majors) |
| Development start | FE 13 / BE 16.5 / QA 8 (37.5 total) | medium-high | 0/13 tasks | TBD | Mostly ready — historical-validation.task-breakdown.md has 11/13 tasks ✅ Actionable or ⚠️ Partially blocked-but-startable now; only 2 are 🚫 fully blocked: Task 1.5 (detail-view pixel QA, gated on REV-6 Figma frame) and Task 2.8 (re-run endpoint, gated on REV-7 product/API decision) |
| Release | — | — | 0/0 | TBD | Planned |
Dependencies & gates
- REV-6 (Design: dedicated comparison/detail Figma frame) gates Task 1.5 (pixel-faithful detail view) — Task 1.2's functional layer is actionable without it.
- REV-7 (product/API decision on the re-run trigger + idempotency guard) gates Task 2.8 (re-run endpoint) — Task 2.4's DELETE half is actionable without it.
- REV-8 (idempotency approach for create: DB index vs. app-level dedup vs. accepted duplicate risk) shapes Task 2.1; team defaults to app-level dedup absent an explicit product call.
- REV-9 (AI squad confirmation of per-org QontakNLP RPM/TPM ceiling; Redis-backed org-level throttling) shapes Task 2.2's production rollout, not its initial build.
- Tasks 2.6 and 2.7 (real API wiring) depend on their respective BE tasks (2.1–2.5) landing first.