Skip to content

docs(7-bm): resolve FLOW-1 command-vs-read; name the continuous-regulation runtime gap#369

Open
xmap wants to merge 1 commit into
mainfrom
worktree-7bm-continuous-regulation
Open

docs(7-bm): resolve FLOW-1 command-vs-read; name the continuous-regulation runtime gap#369
xmap wants to merge 1 commit into
mainfrom
worktree-7bm-continuous-regulation

Conversation

@xmap

@xmap xmap commented Jun 25, 2026

Copy link
Copy Markdown
Owner

What

Resolves the command-vs-read half of 7-BM FLOW-1 and names the continuous-regulation runtime gap precisely. Docs/descriptor only; 7-BM stays a design-phase scaffold (no scenario, no spine code).

Why

The audit named continuous regulation (a setpoint held/ramped during a Run, orthogonal to the scan step-list) the deepest remaining cross-facility gap. This pass settles the part FLOW-1 actually blocks on and pins the rest:

  • Settled: CORA commands the regulated gas-flow / air setpoint. The FlowController presents the earned Regulator Role; a one-shot SetpointStep, not read-back-only.
  • Named, not built: a continuous setpoint program (a hold/ramp held during a Run) has no runtime today, SetpointStep / ControlPort.write are one-shot and the Conductor walks a finite step list. The regulation loop itself stays device/IOC-owned (the Sierra controller runs it); CORA's gap is expressing and observing the program, not hosting the loop. Deferred to a Stage-0 research note + a later gate-reviewed build, exactly as the event-stream axis was for XFEL/XPCS acquisition.

FlowController stays loose (7-BM + i22, short of rule-of-three). FLOW-1 stays live in questions.md / inventory.md for the two remaining parts (FlowController graduation + the setpoint-program runtime).

Files

  • deployments/7-bm/beamline.yaml - FlowController note
  • docs/deployments/7-bm/{model,questions,inventory}.md

Verification

  • pytest tests/unit/deployments/ - 275 passed (incl. test_open_questions_alignment)
  • make docs-build - clean

No new Family / Role / Method / kind (no naming-r3 needed); Regulator already exists, FlowController stays loose.

🤖 Generated with Claude Code

…ation runtime gap

The command-vs-read half of FLOW-1 is settled: CORA COMMANDS the regulated
gas-flow / air setpoint (the FlowController presents the earned Regulator Role;
a one-shot SetpointStep), not read-back-only.

What stays open is split out and named precisely. A continuous setpoint PROGRAM,
a hold or ramp held during a Run while the scan acquires, has no runtime today:
the Conductor walks a finite step list and SetpointStep / ControlPort.write are
one-shot. The regulation loop itself is correctly device/IOC-owned (the Sierra
controller runs its own loop); CORA's gap is expressing and observing the
program, not hosting the loop. This is the deepest cross-facility architectural
gap the device audit named, and 7-BM (flow/combustion) is its grounding case
alongside i11 / XPD thermal. It is deferred to a Stage-0 research note and a
later gate-reviewed build, exactly as the event-stream axis was for XFEL/XPCS
acquisition.

FlowController stays loose (7-BM + I22, short of rule-of-three; not graduated).
FLOW-1 stays live in questions.md / inventory.md for the two remaining parts
(FlowController graduation + the setpoint-program runtime), so the
test_open_questions_alignment guard keeps the question and its placeholders in step.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown

Coverage report

This PR does not seem to contain any modification to coverable code.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant