masarray · masarray · Jul 12, 2026 · Jul 12, 2026 · Jul 12, 2026 · Jul 12, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -4,7 +4,16 @@ All notable ARSVIN changes are documented here using a lightweight Keep a Change
 
 ## Unreleased
 
-_No unreleased changes._
+### Added
+
+- Added a formal Sampled Values standards and evidence research gate covering authoritative source hierarchy, claim levels, OMICRON benchmark findings, licensed-standard gaps, and implementation acceptance rules.
+- Added a conformance and interoperability matrix separating verified, implemented-generic, provisional, unknown, and out-of-scope behavior for Layer-2 SV, 9-2LE-style workflows, IEC 61869-9, timing, scaling, redundancy, and product claims.
+- Added a safe Sampled Values evidence-intake guide for anonymized SCL/PCAP cases, real-device metadata, independent verification, minimum negative tests, and evidence quality levels.
+
+### Changed
+
+- Documentation now prevents unverified profile constants from being promoted directly into production support claims.
+- The SV roadmap now treats StationScout as a system-engineering workflow benchmark and DANEO 400 as the stronger benchmark for SCL-versus-wire verification, merging-unit analysis, timing statistics, orphan detection, triggered recording, and evidence reporting.
 
 ## 0.4.0 — 2026-07-12
 ### Added

diff --git a/docs/index.md b/docs/index.md
@@ -8,6 +8,12 @@ ARSVIN is a focused IEC 61850 Sampled Values engineering suite for Windows. It i
 - [Safety Boundaries](safety-boundaries.md) — laboratory assumptions and responsible use.
 - [Known Limitations](known-limitations.md) — current protocol and runtime boundaries.
 
+## Sampled Values standards and interoperability
+
+- [SV Standards and Evidence Research Gate](sv-research-gate.md) — authoritative source hierarchy, OMICRON benchmark findings, claim levels, and research acceptance rules.
+- [SV Conformance and Interoperability Matrix](sv-evidence-matrix.md) — verified, provisional, unknown, and out-of-scope requirements for 9-2LE-style, IEC 61869-9, and generic Layer-2 SV.
+- `samples/sv-evidence` — safe evidence intake, anonymization rules, metadata, and minimum regression cases for real merging-unit captures.
+
 ## Publisher workflows
 
 - [SV Profile Support](sv-profile-support.md) — current IEC 61850 SV publishing support and limits.
@@ -45,3 +51,4 @@ ARSVIN is a focused IEC 61850 Sampled Values engineering suite for Windows. It i
 - `samples/comtrade` — COMTRADE replay files.
 - `samples/evidence` — sample Markdown evidence reports.
 - `samples/scenarios` — publisher scenario notes and matrices.
+- `samples/sv-evidence` — anonymized or synthetic profile-validation evidence packages.
diff --git a/docs/sv-evidence-matrix.md b/docs/sv-evidence-matrix.md
@@ -0,0 +1,160 @@
+# Sampled Values Conformance and Interoperability Matrix
+
+This matrix is the working control sheet for ARSVIN profile expansion. It records what is known, what remains provisional, and what evidence is required before a behavior becomes part of a supported profile.
+
+## Evidence status vocabulary
+
+| Status | Meaning |
+|---|---|
+| Verified-standard | Confirmed from a licensed applicable standard and recorded with edition and clause. |
+| Verified-official | Confirmed from an official manufacturer or standards-body public document. |
+| Verified-capture | Confirmed from an anonymized real-device SCL/PCAP pair. |
+| Verified-lab | Confirmed by byte-exact or interoperability testing with a real counterpart. |
+| Implemented-generic | Engine supports the mechanism, but no profile-specific compliance claim is made. |
+| Provisional | Candidate behavior awaiting stronger evidence. |
+| Unknown | Not yet researched. |
+| Out-of-scope | Explicitly excluded from the current milestone. |
+
+## Standards source register
+
+| Source | Public edition baseline | Licensed clause review | Owner action |
+|---|---|---:|---|
+| IEC 61850-6 | 2009 + AMD1:2018 + AMD2:2024 CSV | Required | Obtain/access current licensed copy; map SCL requirements. |
+| IEC 61850-7-2 | 2010 + AMD1:2020 CSV | Required | Map abstract SV service semantics. |
+| IEC 61850-9-2 | 2011 + AMD1:2020 CSV | Required | Map Ethernet/APDU encoding and required fields. |
+| IEC/IEEE 61850-9-3 | 2016 | Required for timing claims | Map PTP profile boundaries and terminology. |
+| IEC 61850-10 | 2012 + AMD1:2025 CSV | Required for claim language | Map conformance terminology and test evidence boundaries. |
+| IEC 61869-9 | Current licensed edition to verify | Required | Map preferred variants, scaling, sampling, and dataset requirements. |
+| IEC TR 61850-90-5 | 2012 public catalogue baseline | Deferred | Separate R-SV architecture program. |
+| UCAIug 9-2LE guideline | Revision to identify | Required for 9-2LE claim | Obtain authoritative copy and revision history. |
+
+## Generic Layer-2 SV mapping
+
+| Requirement | Current ARSVIN state | Evidence status | Required next evidence |
+|---|---|---|---|
+| Ethernet EtherType `0x88BA` | Publisher and Subscriber use Layer-2 SV EtherType | Implemented-generic | Clause mapping and byte-exact fixture. |
+| APPID and process-bus header | Implemented | Implemented-generic | Boundary and malformed-frame vectors. |
+| VLAN tag, ID, and priority | Implemented | Implemented-generic | Standard clause mapping and tagged/untagged captures. |
+| `svID` / `smvID` | Parsed and emitted | Implemented-generic | Edition-specific naming and length tests. |
+| Dataset reference | SCL-driven | Implemented-generic | Clause mapping and dataset-resolution fixtures. |
+| `confRev` | Parsed and emitted | Implemented-generic | Change/mismatch behavior tests. |
+| `smpCnt` | Parsed, emitted, monitored, and wrap-tested | Implemented-generic | Profile-specific wrap rules. |
+| `smpSynch` | Compatibility modes implemented | Implemented-generic | Licensed semantics and real synchronized/unsynchronized captures. |
+| `smpRate` / `smpMod` | Parsed and emitted | Implemented-generic | Edition/profile-specific presence and interpretation rules. |
+| Multiple ASDUs per frame | Engine supports `1..8`; UI/docs historically emphasize `1/2/4/8` | Implemented-generic | Profile-specific allowed values and `nofASDU=6` evidence. |
+| Dataset-driven payload | Multiple integer, float, quality, time, enum, bit-string, and string kinds supported | Implemented-generic | Validate permitted SV bTypes, widths, order, and nesting. |
+| Unknown payload element | Preflight can reject unsupported layouts | Implemented-generic | Compatible-mode raw preservation design. |
+| PCAP read/write | Implemented and regression-tested | Implemented-generic | Golden external captures. |
+| Timing health | Local TX and receiver statistics exist | Implemented-generic | Define measurement method and limitations against DANEO-like outcomes. |
+
+## 9-2LE-style profile research
+
+| Topic | Current position | Status | Exit criterion |
+|---|---|---|---|
+| Profile name and revision | `9-2LE-style` only | Provisional | Identify authoritative guideline and revision. |
+| 4 current + 4 voltage layout | Current primary workflow | Provisional | Guideline clause plus golden dataset and captures. |
+| Value representation | Current workflow uses SCL-derived layout | Provisional | Byte-level guideline verification. |
+| Quality ordering and encoding | Engine supports quality fields | Provisional | Guideline vectors and device capture. |
+| Protection sampling variant | Product assumption exists | Provisional | Standard/guideline source plus 50 Hz and 60 Hz captures. |
+| High-rate variant | Roadmap candidate | Provisional | Standard/guideline source and performance fixture. |
+| Nominal-frequency behavior | Not yet formalized as profile evidence | Unknown | Verify formula, rate, wrap, and frame-rate behavior. |
+| Scaling convention | Not yet provenance-aware | Unknown | Verify raw-to-secondary convention and expose source. |
+| Publisher interoperability | Generic output works | Provisional | At least one real relay/MU test per claimed variant. |
+| Subscriber detection | SCL-assisted decode exists | Provisional | Confidence classifier and known/unknown/mismatch vectors. |
+
+## IEC 61869-9 profile research
+
+| Topic | Current position | Status | Exit criterion |
+|---|---|---|---|
+| Preferred variant catalogue | Not implemented | Unknown | Licensed clause matrix. |
+| Configurable dataset behavior | Generic SCL engine can support it | Provisional | Standard requirement and real SCL evidence. |
+| Protection-oriented variant | Roadmap candidate | Provisional | Licensed rates, packing, payload, and timing requirements. |
+| Measurement/high-rate variant | Roadmap candidate | Provisional | Licensed requirements and sustained-throughput tests. |
+| Sampling basis | Engine supports SCL `smpRate`/`smpMod` | Provisional | Verify samples-per-second versus samples-per-cycle semantics. |
+| `nofASDU` values | Engine mechanically supports `1..8` | Provisional | Profile-specific normative/recommended values. |
+| Dataset shape and channel order | Must not be assumed fixed 4I+4V | Provisional | Standard plus vendor SCL matrix. |
+| Scaling and units | Not yet profile-provenance-aware | Unknown | Standard clause, SCL source, and real values. |
+| Synchronization | Generic `smpSynch` and PTP tooling exist | Provisional | Standard linkage and capture evidence. |
+| Public support claim | Not supported yet | Unknown | All mandatory rows reach verified-standard and fixture/device evidence. |
+
+## OMICRON-derived product benchmark matrix
+
+This table records product outcomes, not protocol requirements.
+
+| Benchmark outcome | StationScout | DANEO 400 | ARSVIN target |
+|---|---:|---:|---|
+| SCL-first system view | Strong | Strong | Import streams and show expected-versus-observed state. |
+| Configuration versus live comparison | Strong | Strong | Side-by-side profile, addressing, dataset, and timing mismatch. |
+| Signal tracing | Strong | Partial for SV analysis | Trace SCL stream to observed stream and decoded channels. |
+| Orphan detection | Not primary public emphasis | Explicit | List observed SV streams absent from SCL and expected streams absent from traffic. |
+| MU parameter verification | Not primary public emphasis | Explicit | Validate stream fields, rate, packing, payload, synchronization, and quality. |
+| Analog input versus SV output | Outside normal ARSVIN PC-only scope | Explicit | Future import/correlation path; do not imply calibrated measurement. |
+| Packet interval/delay statistics | Not primary public emphasis | Explicit | Add interval, jitter, gaps, delay-input limitations, histograms, and evidence export. |
+| Triggered recording | Test-case oriented | Explicit | Add software capture trigger rules after profile detector is stable. |
+| PRP/HSR verification | Not primary public emphasis | Explicit | Separate redundancy milestone after base SV profiles. |
+| Reusable tests and reports | Explicit | Explicit | Machine-readable expected-profile fixtures and repeatable reports. |
+
+## Fast-track implementation gates
+
+### Gate 0 — Research register
+
+Complete when:
+
+- source editions are recorded,
+- licensed-source gaps are visible,
+- every proposed profile constant is marked verified or provisional,
+- and no unsupported public claim is introduced.
+
+### Gate 1 — Profile infrastructure
+
+Complete when:
+
+- observed facts are independent of WPF and Npcap,
+- profile definitions carry source/evidence metadata,
+- confidence scoring explains matches and conflicts,
+- and unknown streams remain observable.
+
+### Gate 2 — Formalized 9-2LE-style support
+
+Complete when:
+
+- the authoritative guideline revision is reviewed,
+- 50 Hz and 60 Hz fixtures exist where applicable,
+- byte-exact Publisher and Subscriber tests pass,
+- scaling provenance is visible,
+- and at least one real counterpart has been exercised.
+
+### Gate 3 — IEC 61869-9 protection support
+
+Complete when:
+
+- licensed clauses are mapped,
+- preferred variant parameters are frozen,
+- SCL and PCAP fixtures pass,
+- real-device evidence exists,
+- and performance remains stable.
+
+### Gate 4 — High-rate and redundancy support
+
+Complete when:
+
+- sustained throughput is benchmarked,
+- queues are bounded,
+- UI rendering is decoupled from packet rate,
+- PRP/HSR behavior has dedicated evidence,
+- and no false sample-gap diagnosis occurs on clean captures.
+
+## Definition of public `Supported`
+
+A profile can be labeled supported only when all are true:
+
+- standard/guideline edition and clause register complete,
+- deterministic fixtures committed,
+- byte-exact generation test passing,
+- subscriber decode and detection tests passing,
+- mismatch and malformed tests passing,
+- scaling source explicit,
+- performance baseline documented,
+- real-device evidence recorded,
+- support matrix updated,
+- and claim language reviewed against IEC 61850-10 boundaries.