secs-gem/COMPLIANCE.md

SECS/GEM Compliance Audit

An honest, per-capability accounting of which parts of SEMI E5 (SECS-II), E30 (GEM), and E37 (HSMS) this codebase implements, which parts it implements partially, and which parts it does not implement at all.

TL;DR. This is not a 100%-conformant GEM equipment. It implements every GEM Fundamental capability except formal Documentation (S1F19–F22) and the S9 error stream, plus most of the high-value GEM Additional capabilities (Dynamic Event Reporting, Alarms, EC, Clock, Remote Control, Process Programs, Terminal). Spooling, Limits Monitoring, Trace Data Collection, Multi-block, and Material Movement are intentionally not yet implemented. See the per-capability tables below for spec section refs.

Legend:

• ✅ Full — message/feature implemented; round-tripped in tests.
• 🟡 Partial — implemented in the demo path but with documented limitations.
• ⬜ Not implemented — intentionally out of scope for this iteration.

---

1. E37 — HSMS transport

Item	Status	Spec ref	Notes
TCP transport	✅	E37 §6	hsms::Connection (Asio).
4-byte length prefix + 10-byte header	✅	E37 §8.2	hsms::Frame::encode/decode.
Session ID, byte2, byte3, PType, SType, system-bytes	✅	E37 §8.3	hsms::Header.
Select.req / .rsp	✅	E37 §7.2	SType 1/2; SelectStatus enum (0–3).
Deselect.req / .rsp	✅	E37 §7.4	SType 3/4; DeselectStatus enum (0–2).
Linktest.req / .rsp	✅	E37 §7.5	SType 5/6; periodic interval configurable.
Separate.req	✅	E37 §7.6	SType 9; graceful close after flush.
Reject.req	🟡	E37 §7.7	Emitted on data-while-NOT-SELECTED with reason 4; we don't yet reject on PType/SType-not-supported.
Connection state machine NOT-CONNECTED → NOT-SELECTED → SELECTED	✅	E37 §6.3	Both Active and Passive modes.
T3 reply timeout	✅	E37 §10	per-transaction steady_timer.
T5 connect separation timeout	✅	E37 §10	Client::schedule_retry.
T6 control transaction timeout	✅	E37 §10	single slot (no concurrent control transactions).
T7 not-selected timeout (passive)	✅	E37 §10	armed on connect / on Deselect.req.
T8 intercharacter timeout	✅	E37 §10	bounds the payload read after length prefix.
Multi-host / multi-session	⬜	E37 §6	single SELECTED session at a time.
HSMS-SS vs HSMS-GS	⬜	E37 §11	implemented HSMS-SS only.

---

2. E5 — SECS-II encoding

Item	Status	Spec ref	Notes
Format byte + 1/2/3 length bytes	✅	E5 §9	secs2::encode_into.
List (L)	✅	E5 §9.3	recursive.
ASCII (A)	✅	E5 §9.5
Binary (B)	✅	E5 §9.5
Boolean (BOOLEAN)	✅	E5 §9.5
U1, U2, U4, U8 (big-endian)	✅	E5 §9.5
I1, I2, I4, I8 (big-endian, two's complement)	✅	E5 §9.5
F4, F8 (IEEE 754 big-endian)	✅	E5 §9.5	bit-cast round-trip.
JIS-8, C2 (Unicode)	⬜	E5 §9.5	rarely used in modern fabs.
SML text rendering	✅	E5 Annex	secs2::to_sml.

---

3. E30 — GEM Fundamental capabilities (§5.2)

Fundamental Capability	Status	Spec ref	Messages	Notes
State models	✅	E30 §6.2	—	Control state machine (5 states) + HSMS comm state. Equipment-processing state is left to the equipment app.
Equipment Processing States	⬜	E30 §6.3	—	Standard says equipment may model these; tool-specific. Plug-in point not yet exposed.
Host-Initiated S1F13/F14 scenario	✅	E30 §6.5	S1F13/F14
Event Notification	✅	E30 §6.6	S6F11/F12	Equipment-initiated, host-acknowledged.
On-Line Identification	✅	E30 §6.7	S1F1/F2	MDLN + SOFTREV.
Error Messages	✅	E30 §6.9	S9F*	HSMS Reject.req at the transport level; S9F1/F3/F5/F7/F9/F11/F13 in the SECS-II catalog. Auto-emitted by Connection on every documented condition: S9F3 (unknown stream) and S9F5 (unknown function) via the server's Router-aware wrapper using the new Connection::current_header() accessor; S9F7 (illegal data) when a peer body fails SECS-II decode; S9F9 (T3 timeout) reconstructing the original outgoing MHEAD; S9F11 (data too long) with a synthesized MHEAD whose first 4 bytes are the offending length prefix. S9F1 / S9F13 are catalog-only — no documented condition currently triggers them in our codebase.
Documentation	✅	E30 §6.10	S1F19/F20, S1F21/F22	Equipment self-reports its GEM-compliance level via S1F20 (SOFTREV, EQPTYP, list of (CCODE, CDESC) per E30 Appendix A) and its DVID namelist via S1F22. Both populated from data/equipment.yaml.
Control (Operator-Initiated)	✅	E30 §6.2	—	ControlStateMachine::operator_online/offline/local/remote.

---

4. E30 — GEM Additional capabilities (§5.3)

Additional Capability	Status	Spec ref	Messages	Notes
Establish Communications	✅	E30 §6.5	S1F13/F14	Both directions modeled; COMMACK enum.
Dynamic Event Report Configuration	✅	E30 §6.6	S2F33/F34, S2F35/F36, S2F37/F38	Full Define-Report / Link-Event / Enable-Event pipeline with all four ack enums.
Variable Data Collection	✅	E30 §6.11	—	DVID table; values resolvable via vid_value.
Trace Data Collection	✅	E30 §6.12	S2F23/F24, S6F1/F2	TraceStore keeps an active TRID -> TraceConfig map; S2F23 validates each SVID exists and registers the request (TIAACK=4 on unknown VID). S6F1 round-trips with TRID/SMPLN/STIME/values. The periodic sampling timer that turns a TraceConfig into S6F1 emissions is intentionally left to the application (vendors typically already have their own scheduler).
Status Data Collection	✅	E30 §6.13	S1F3/F4, S1F11/F12
Alarm Management	✅	E30 §6.14	S5F1/F2, S5F3/F4, S5F5/F6, S5F7/F8	Full set: emit, ack, enable/disable, list directory, list enabled. ALCD bit-7 set/cleared, lower-7 category.
Remote Control	✅	E30 §6.15	S2F41/F42	Full HCACK 7-value enum + per-parameter CPACKs.
Equipment Constants	✅	E30 §6.16	S2F13/F14, S2F15/F16, S2F29/F30	EAC range validation against min/max is NOT performed — set always accepts a known ECID; this would reject out-of-range with EAC=4 in a conformant equipment.
Process Program Management	🟡	E30 §6.17	S7F3/F4, S7F5/F6, S7F19/F20	Unformatted PP send/request/list. No E42 enhanced PP, no S7F23/F24/F25/F26 PP verify, no PPID validation. PPBODY treated as raw bytes (Binary item).
Material Movement	⬜	E30 §6.18	S3F*, E40	Tied to E40 carrier handling; separate standard.
Equipment Terminal Services	🟡	E30 §6.19	S10F1/F2, S10F3/F4	Single-line only. S10F5/F6 multi-block, S10F7 broadcast not implemented.
Clock	✅	E30 §6.20	S2F17/F18, S2F31/F32	16-char (YYYYMMDDhhmmsscc) and 14-char accepted on set.
Limits Monitoring	✅	E30 §6.21	S2F45/F46, S2F47/F48	LimitMonitorStore keyed by VID with multiple LimitDefinition (LIMITID + upper/lower as arbitrary Items). Server S2F45 validates each VID exists (returns VLAACK=4 otherwise) and stores definitions; S2F47 returns them. Crossing-detection + CEID emission is intentionally left to the application's set_value path (E30 §6.21 is permissive about how the equipment detects crossings).
Spooling	🟡	E30 §6.22	S2F43/F44, S6F23/F24	Spoolable-streams config (S2F43/F44) + spool request transmit/purge (S6F23/F24) implemented; SpoolStore queues equipment-initiated primaries when the host is unreachable or force-spool is on, drains FIFO on S6F23 transmit, purges on S6F23 purge. Not implemented: S6F25/F26 spooled-data-ready notification, automatic spool activation tied to HSMS NOT-SELECTED, persistent (on-disk) spool.
Control	✅	E30 §6.2	—	See Fundamental.

---

5. Message coverage matrix

Pair	Direction	Status	Implemented in	Tested
S1F1 / S1F2	H↔E	✅	messages.hpp	✅ round-trip
S1F3 / S1F4	H→E	✅	messages.hpp	✅ round-trip
S1F11 / S1F12	H→E	✅	messages.hpp	✅ round-trip
S1F13 / S1F14	H↔E	✅	messages.hpp	✅ in demo
S1F15 / S1F16	H→E	✅	messages.hpp	✅ in demo
S1F17 / S1F18	H→E	✅	messages.hpp	✅ in demo
S1F19 / S1F20	H→E	✅	catalog	✅ round-trip + demo
S1F21 / S1F22	H→E	✅	catalog	✅ round-trip + demo
S2F13 / S2F14	H→E	✅	messages.hpp	✅ in demo
S2F15 / S2F16	H→E	✅	messages.hpp	✅ round-trip
S2F17 / S2F18	H→E	✅	messages.hpp	✅ round-trip
S2F23 / S2F24	H→E	✅	catalog	✅ round-trip
S2F29 / S2F30	H→E	✅	messages.hpp	✅ in demo
S2F31 / S2F32	H→E	✅	messages.hpp	✅ in demo
S2F33 / S2F34	H→E	✅	messages.hpp	✅ round-trip
S2F35 / S2F36	H→E	✅	messages.hpp	✅ round-trip
S2F37 / S2F38	H→E	✅	messages.hpp	✅ round-trip
S2F41 / S2F42	H→E	✅	messages.hpp	✅ round-trip
S2F45 / S2F46	H→E	✅	catalog	✅ round-trip
S2F47 / S2F48	H→E	✅	catalog	✅ round-trip
S5F1 / S5F2	E→H	✅	messages.hpp	✅ round-trip
S5F3 / S5F4	H→E	✅	messages.hpp	✅ round-trip
S5F5 / S5F6	H→E	✅	messages.hpp	✅ in demo
S5F7 / S5F8	H→E	✅	catalog	✅ round-trip
S6F1 / S6F2	E→H	✅	catalog	✅ round-trip
S6F5 / S6F6	H↔E	⬜	—	multi-block
S6F7 / S6F8	H↔E	⬜	—	multi-block
S6F11 / S6F12	E→H	✅	messages.hpp	✅ round-trip + demo
S6F15 / S6F16	H→E	⬜	—	event report request
S2F43 / S2F44	H→E	✅	catalog	✅ round-trip + demo
S6F23 / S6F24	H→E	✅	catalog	✅ round-trip + demo
S6F25 / S6F26	spool	⬜	—	spool-data-ready notification
S7F3 / S7F4	H→E	✅	messages.hpp	✅ round-trip
S7F5 / S7F6	H→E	✅	messages.hpp	✅ in demo
S7F19 / S7F20	H→E	✅	messages.hpp	✅ round-trip + demo
S7F23–F26	H↔E	⬜	—	enhanced PP
S9F1	E↔H	✅	catalog	✅ round-trip
S9F3	E↔H	✅	catalog + Router wrapper	✅ round-trip + auto-emitted on unknown stream
S9F5	E↔H	✅	catalog + Router wrapper	✅ round-trip + auto-emitted on unknown function
S9F7	E↔H	✅	catalog + Connection	✅ round-trip + auto-emitted on body decode
S9F9	E↔H	✅	catalog + Connection	✅ round-trip + auto-emitted on T3 timeout
S9F11	E↔H	✅	catalog + Connection	✅ round-trip + auto-emitted on oversized frame
S9F13	E↔H	✅	catalog	✅ round-trip
S10F1 / S10F2	H→E	✅	messages.hpp	✅ in demo
S10F3 / S10F4	E→H	✅	messages.hpp	✅ round-trip + demo
S10F5 / S10F6	H→E	⬜	—	multi-line

---

6. Demo evidence

The two-container demo (docker compose up --no-deps server client) exercises this concrete sequence end-to-end:

1. TCP connect → Select.req → Select.rsp(Ok) → SELECTED on both sides.
2. S1F13/S1F14 Establish Comms.
3. S1F17/S1F18 Request Online; control state transitions HostOffline → AttemptOnline → OnlineRemote.
4. Server pushes S10F3 welcome → host acks with S10F4.
5. S1F11/S1F12 SVID namelist discovery → S1F3/S1F4 SVID read.
6. S2F29/S2F30 EC namelist → S2F13/S2F14 EC read.
7. S2F17/S2F18 clock read.
8. S2F33/S2F34 Define Report 1000 over the 3 SVIDs (DRACK=0).
9. S2F35/S2F36 Link CEIDs 200 and 300 to Report 1000 (LRACK=0).
10. S2F37/S2F38 Enable CEIDs 200, 300 (ERACK=0).
11. S2F41/S2F42 host command START (HCACK=0) → server emits S6F11(CEID=300) carrying the linked Report 1000 → host acks S6F12.
12. S5F5/S5F6 list alarm directory.
13. S5F3/S5F4 enable alarm 1 (ACKC5=0).
14. S2F41/S2F42 host command FAULT (HCACK=0) → server emits S5F1 (ALCD=0x84, set + cat 4) → host acks S5F2; server also emits S6F11(CEID=200).
15. S7F19/S7F20 recipe list, S7F5/S7F6 fetch RECIPE-A.
16. S10F1/S10F2 host → equipment terminal display.
17. S1F15/S1F16 Request Offline; control state goes back to HostOffline. (CEID 100 ControlStateChanged emission is correctly suppressed because the host never enabled CEID 100 — this is the correct GEM behavior.)
18. Separate.req → clean close on both sides.

Unit tests: 63 cases / 278 assertions pass (docker compose run --rm tests).

---

7. What it would take to claim "100% GEM-compliant"

The honest list, in priority order:

1. Finish spooling: S6F25/F26 spooled-data-ready notification, plus automatic activation when HSMS goes NOT-SELECTED (and automatic notification on re-SELECT) so the host doesn't have to manually flip the test-only force_spool flag. Optional: persistent on-disk spool so equipment restarts don't lose queued events.
2. (done in this revision): S9F3/F5/F11 — all S9 conditions now auto-emit.
3. (done in this revision): S1F19/F20 + S1F21/F22 — Documentation.
4. Implement EC range validation in set_equipment_constant_value so out-of-range sets return EAC=4 instead of being silently accepted.
5. Implement Limits Monitoring (S2F45–F48, S6F45/F46) if the target equipment publishes monitored variables.
6. Implement Trace Data Collection (S2F23/F24, S6F1/F2).
7. Implement S5F7/F8 list-enabled-alarms.
8. Implement multi-block transfers (S6F5/F6, S6F7/F8).
9. Implement equipment processing state model with operator hooks (E30 §6.3) — the abstract model is in E30 but the concrete states are equipment-specific.
10. Run the implementation against a real conformance test generator (Layer 4 of the implementation plan) on a representative tool.

After all of the above, a GEM Reference Test specification (RTS) review would still be needed before any "GEM compliant" marketing claim could be honestly made.