docs(splat-native): §6 FMA-litmus customer narrative — splat-native ultrasound as the load-path customer of Phase 8 FMA hydration#34
Merged
Conversation
Cross-session companion to the lance-graph splat-native-ultrasound-v1 canonical plan. Names splat-native ultrasound as the explicit downstream customer of OGAR PR #30 §6 FMA bones-rendering litmus + Phase 8 FMA hydration. The §6 litmus transitions from "demo target" to "load path" — splat-native is the contemporary customer of the FMA atlas substrate at the sub-millisecond HHTL traversal claim. The SaMD Class IIa certification track uses ADR-022 (The Firewall) + KnowableFromStore (PR #25/#31) + commit_event audit (callcenter PR #467) as the audit-controls evidence base — no new OGAR architecture for SaMD; only documentation that names what's already true. No code. Spec/narrative only.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ae92159d8e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+245
to
+246
| - **IVD-MDR Rule 11** — EU SaMD risk classification (Class IIa for | ||
| diagnostic-software with non-critical decision support). |
There was a problem hiding this comment.
For the EU SaMD path described here, Rule 11/Class IIa is an MDR Annex VIII classification, not an “IVD-MDR”/IVDR one; IVDR is for in-vitro diagnostics and uses A–D risk classes. Because this section is the regulatory cross-reference future D-SPLAT-14 work will follow, leaving the wrong regulation name can send the certification/evidence plan down the wrong rule set for the stated ultrasound SaMD device.
Useful? React with 👍 / 👎.
AdaWorldAPI
added a commit
that referenced
this pull request
Jun 5, 2026
…ex VIII Rule 11) Follow-up to PR #34 (merged). Addresses one codex P2 finding. ## Fix — Regulatory basis for Rule 11 / Class IIa (codex P2) The §4 standards / vocabularies index cited "IVD-MDR Rule 11" as the EU SaMD risk classification for splat-native ultrasound. This is the wrong regulation. IVDR (the In-Vitro Diagnostic Regulation 2017/746) governs IN-VITRO DIAGNOSTICS (e.g., blood tests, immunoassays) and uses A-D risk classes. Ultrasound SaMD is regulated under the MDR (Medical Device Regulation 2017/745); the applicable classification rule is **MDR Annex VIII Rule 11**. Because this section is the regulatory cross-reference future D-SPLAT-14 (SaMD documentation track) work will follow, leaving the wrong regulation name would send the certification/evidence plan down the wrong rule set for the stated ultrasound device. Fix: corrects the IVD-MDR entry to MDR Annex VIII Rule 11 and adds a half-sentence call-out explaining that IVDR is a separate framework that does NOT govern this device. The Class IIa classification claim is unchanged (just attached to the correct regulation now). Companion fix in the lance-graph canonical plan (PR follow-up to #471) which had the same wrong citation in the D-SPLAT-14 row of both the plan body and the STATUS_BOARD. ## What's NOT in this PR - Source code: still none. Narrative-only doc. - The other regulatory standards (IEC 62366 usability, IEC 80001 risk, ISO 14971 risk-management) are unchanged. - The Class IIa classification claim itself is unchanged (it was correct; only the regulation citation was wrong). ## Test plan - [x] Codex P2 (regulatory basis) — IVD-MDR → MDR Annex VIII Rule 11. - [x] Cross-grep: no remaining IVD-?MDR references in this file. - [ ] Codex re-review on this PR.
This was referenced Jun 5, 2026
AdaWorldAPI
added a commit
that referenced
this pull request
Jun 5, 2026
…fixes docs(splat-native): address review feedback on #34 (IVD-MDR → MDR Annex VIII Rule 11)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
OGAR-side companion to the cross-workspace splat-native-ultrasound-v1 integration plan (canonical at
lance-graph/.claude/plans/splat-native-ultrasound-v1.md). This PR ships one new doc that names splat-native ultrasound as the explicit customer of OGAR PR #30 §6 FMA-litmus + Phase 8 FMA hydration.What this PR ships
docs/SPLAT-NATIVE-CUSTOMER.md(~260 lines).No code. Narrative + cross-reference doc only.
TL;DR (from the doc)
OGAR PR #30 §6 names FMA bones-rendering as the architectural litmus:
Splat-native ultrasound is the customer of that litmus. A CPU-only Gaussian-splat ultrasound SaMD pipeline (the canonical plan in lance-graph) needs exactly this query at exactly this latency, on exactly this corpus (~75K FMA classes, ~2.1M relationships), for the splat-to-splat registration loop (live splat volume ↔ FMA atlas splat volume, via Σ-sandwich Mahalanobis ICP).
Three high-leverage takeaways for OGAR:
Phase 8 from PR docs: RDF / OWL ↔ OGAR alignment — Morris triad + brutal-upgrade vision #30 is no longer abstract — it has a concrete customer waiting downstream. The TTL produced by Phase 8 is consumed by lance-graph D-SPLAT-8 (FMA atlas hydrator) which emits typed
FmaEntitySoA + a pre-computed atlas splat volume (~150M Gaussians full body, ~5 GB compressed via the SH-aware palette extension in D-SPLAT-4).ADR-022 (The Firewall) is the SaMD evidence base. The splat-native arc lands as a SaMD Class IIa device. The audit-controls + access-controls evidence the regulator wants IS the ADR-022 firewall discipline +
LanceMembrane::commit_eventaudit chain +KnowableFromStoreregistry that PRs feat: ogar-knowable-from — trait-mediated §10.3 producer seam (ADR-010 OGAR-side close) #25/docs: THE-FIREWALL — absolute inner/outer boundary, no serialization in hot path (ADR-022) #26/docs: HEALTHCARE-TRANSCODING — OGAR for HIPAA/healthcare (FHIR-grounded, label-free) #28/fix: #25 P2 canonical-identity keying + docstring back-refs + F-1 attestation #31 already shipped. No new OGAR architecture for SaMD certification — only documentation that names what's already true.The §6 FMA litmus is reframed from "demo target" to "load path." The sub-millisecond HHTL traversal claim IS the splat-native frame-rate budget's gating threshold (<150ms/frame total budget; sub-ms NiblePath traversal + <10ms atlas-region paging + <100ms ICP + <33ms render).
Composed litmus (the SaMD acceptance gate)
What this arc adds to OGAR's roadmap
Confirms (no new work)
KnowableFromStoreregistry-assigns pattern + canonical-identity keying are splat-native-readySurfaces (small, additive; sized within existing Phase 8 budget)
vocab/fma-alignment.ttl(FMA ↔ OGARClasscorrespondence)vocab/ogit-fma.ttldeclaring FMA-specific OGIT predicates (isA,partOf,innervates,supplies)Reframes (zero new work; framing only)
Three FINDINGs (verified against existing OGAR + lance-graph state)
KnowableFromStoreregistry-assigns pattern (PR feat: ogar-knowable-from — trait-mediated §10.3 producer seam (ADR-010 OGAR-side close) #25/fix: #25 P2 canonical-identity keying + docstring back-refs + F-1 attestation #31) is splat-native-ready. Splat ingest callsregister("ogit-medcare/ultrasound_ingest", Some(ddl_hint))and gets back a monotonicu64. Per amended C-2 in MedCare PR #162.HEALTHCARE-TRANSCODING.md §3.3production-instance reference absorbs the splat use case without modification.Three CONJECTUREs (need OGAR-session confirmation)
vocab/ogit-fma.ttlminimum predicate set sufficient for registration (4 named; defer to D-SPLAT-9 review).Action items for OGAR session (when Phase 8 work opens)
vocab/fma-alignment.ttlandvocab/ogit-fma.ttlTest plan
Cross-PR coordination
This is one of four coordinated PRs for the splat-native-ultrasound-v1 cross-workspace plan. All four reference each other; lance-graph is canonical.
lance-graphclaude/splat-native-ultrasound-v1ndarrayclaude/splat-native-ultrasound-v1MedCare-rsclaude/splat-native-ultrasound-v1OGARclaude/splat-native-customerTriggered by user-supplied architecture diagrams (English 6-stage technical + German business-facing). No source code in any of the four PRs — design-spec / handover only.
Authored by session
claude/lance-graph-ontology-review-Pyry3.