Dashboard Is Not Control

Dashboard is not control. A dashboard reads a stored or streamed value — a historian tag, a CMMS count, a health score, an alarm state already held in the system that owns it — and renders that value as a tile, a trend, or a color. Refreshing the screen, filtering the rows, changing the time range, or drilling into a point repeats the read. It does not write a setpoint, start or stop equipment, open or close a valve, isolate a circuit, or create a work order. Control is that write. It requires a named actor and a system that already accepts the command. Sync may surface the read beside a Decision Case. Surfacing does not attach write credentials to the case. A named human decides. Authorized execution systems write the work order or the isolation. Direct plant execute stays off.

A live screen looks like a station that can act. The number moves. The color changes. The meeting treats the person looking at the screen as if looking were commanding. It is not. The screen updated because a query returned a newer value. The plant may have changed, or it may not have. The color does not say which. The process can move on its own. A sensor can fail. Someone else can send a command from a system that accepts writes. The dashboard does not record which of those happened. It displays a value.

Sync keeps that split on the signed-in Decision Case. A signed-in user completes the case in a fixed order: Question, Evidence, Recommendation, Human decision, Action, Verification, and Learning. Orville Davis states that order in Field Manual v0. The manuals index lives at /manuals. This essay is why a rendered value cannot be read as control of the plant. The Action chapter records work intent after a named decision. The intent is still not the write.

A dashboard is a read

The display path has four steps. A source holds a value: a historian tag, a CMMS count, a calculated health score, or an alarm bit in the system that owns it. A query or a subscription copies the current value out. A renderer maps that copy to a tile, a line, or a color. A person looks at the renderer.

None of those steps sends a command back. The system that stores the displayed value and the system that can change the asset are often different systems. When they are the same system, the display path and the command path are still different calls. One returns a value. The other accepts a write only if the caller holds the authority and the credentials that system requires.

Refresh, filter, sort, time range, drill-down, export, and screenshot stay on the read path. Each one asks for values again or rearranges values already returned. None of them is a setpoint.

Map Is Not Terrain already treats a dashboard as a compressed representation, not the operating reality. This essay is the call underneath the picture. The picture is a copy of a value at a time. It is not a command, and it is not the asset.

A command on the same screen is a different call

Some operator stations draw a trend and also offer a setpoint field, a start, or a stop. The trend is the read. The field is a write only when it is wired to a system that accepts the command, and only when the person using it holds the operating authority and the credentials that system requires. Sharing a screen does not merge the two calls. A chart placed next to a button is not the button.

Sync does not provide that write. This edition does not describe plant execute. It does not describe Sync issuing setpoints, permissives, starts, stops, isolations, or work orders. A dashboard surfaced beside a Decision Case is the read. The case can store a copy of what the screen showed. That copy does not become a control station.

A value rendered on the screen is not a command. Sync does not hold the credentials that would send one.

Control is the write

Control authority over equipment state or maintenance work sits with the people and the systems that already hold it. A setpoint, a start, a stop, or a permissive goes through the control system the site already uses, under the people who hold operating authority. Isolation is the site practice for taking equipment out of service. A work order is written in the customer CMMS or EAM, or the equivalent system of record.

A tile does not hold those authorities by being visible. A trend does not hold them by updating. A Decision Case that stores a reading from the dashboard stores a record of a display. Storage is not a command.

Sync may show the dashboard. A named human decides. Authorized execution systems write the work order or the isolation.

Coverage Is Not Control names the watched set: sensors, CMMS rows, patrol routes, and the tiles on a screen. Coverage is what is watched. It is not permission to order work, isolate equipment, or change the plant. This essay is the split under the tile. The tile is a read of a value. Control is a write. A current tile can exist with no write path from that screen. A write path can exist in the control system while the dashboard only reads.

Color does not name the actor

A green tile, a quiet board, or a red tile is a rendering rule applied to a value. The same color can come from the process sitting inside the rule, from a sensor that stopped reporting, or from a command someone else already sent. The color does not name who sent a write. It does not say whether any write was sent.

A quiet dashboard is not clearance. Silence Is Not Clearance states that boundary for no reply, no alert, and an empty inbox. No alarm on a dashboard is not approval to run, and it is not evidence that a command is holding the asset.

Seeing the color is not a Decision Case and not a decision. Dashboard Is Not Decision keeps the display off the named act. The screen can change while no person has accepted, rejected, escalated, or returned a recommendation. This essay is the next boundary on the same screen. The display is not the decision, and the display is not the write.

A missing tile is not a measured zero. Blank Is Not Zero keeps an empty field from being stored as a measurement of none. An empty cell means the read did not return a value. It does not mean the asset was commanded to zero, and it does not mean the asset is under control.

What the Decision Case may store

Evidence may cite what a display held when the source and the time are named. That citation is a record of a read. It is not a record that Sync commanded the asset. The Evidence chapter publishes that step. Stage-1 evidence is the record held on the case. A live connector that pulls historian or control-system tags sits outside this edition. Simulated or seeded telemetry and assets are practice records. They are not customer plant control, and a practice dashboard is not a command.

A recommendation may say investigate because of what the display showed. The proposal states a next action, the evidence it uses, and the uncertainty it leaves in view. Drafting it does not change the plant. Recommend Is Not Authorize states that split. A named person accepts, rejects, escalates, or returns. The Human decision chapter publishes the act. The act records who decided. It still does not write a setpoint, isolate equipment, or create a work order.

If the named person approves work, the case may store the intent. Action Is Not Execution keeps that intent off the plant. ACTION stays locked until authorized execution systems write the work order or isolate the equipment. A refreshed dashboard does not unlock it. A tile is not a work order. Sync does not write the work order. The systems that already hold execution authority write the work order or the isolation.

Verification asks whether the authorized action did what the decision named. The Verification chapter publishes the check. Verification Is Not Optional states the gate: the case stays open until the check is recorded. The check is the observed outcome of the write, not a later color on the screen. A tile that changes after someone else operated the plant is not Sync controlling the plant, and it is not the check.

Learning Requires a Verified Outcome states what a later case may inherit. Learning keeps the closed case. It does not keep a color change, a refreshed tile, or a dashboard that was never tied to an authorized write. A later question that cites a green board as if the plant were under control is citing a rendering rule.

A telemetry feed, a precise residual, or a calibration can be drawn on the same screen. Drawing it does not make the screen a command, and it does not make the drawing truth, accuracy, or validation. Telemetry Is Not Truth keeps a stream of tags from being stored as the plant decision. Accuracy Is Not Precision keeps repeatability from being stored as truth relative to that decision. Calibration Is Not Validation keeps a small historical residual from being stored as fitness for the decision. The residual can sit in a tile. The tile is still a read.

Where the public statement lives

Field Manual v0 is the public contents of this loop. Start at the manuals index or open Sync Field Manual directly. Evidence may hold what a display showed. Action may hold work intent after a named decision. Neither step is a write to the plant. The Honesty boundaries keep this edition from treating a rendered value as plant control. Later editions can deepen a chapter. The spine stays in this order.

What this article is not claiming

This is an essay about the Decision Case order, not a customer case study. It names no plant, states no savings figure, and claims no prevented failure. It states no OEM limit and no operating threshold. It does not claim that a dashboard writes a setpoint, starts or stops equipment, isolates a circuit, creates a work order, or controls a plant. It does not claim that Sync executes plant work.

Stage-1 readiness means a signed-in user can complete the Decision Case — question, evidence, recommendation, human decision, action, verification, and learning — and Field Manual v0 describes that journey. This edition does not describe plant execute, a live connector tag pull, SMTP invite delivery, or automatic revocation of access on expiry as live. It does not describe Sync writing work orders, isolating equipment, or controlling the plant. Simulated or seeded telemetry and assets are practice records. They are not live plant results. Self-guided onboarding is not claimed as a live product path.

The series continues with Green Is Not Go, on why a green tile, a green status, a green health score, or an all-green dashboard is a display of a stored or computed value under the thresholds someone chose, and why that color is not permission to run, clear, start, or leave equipment in service. Companion reading: Telemetry Is Not Truth on why a stream of tags, sensors, historians, SCADA, and CMMS counters is not truth for the plant decision, Accuracy Is Not Precision on why repeatability is not truth relative to the plant decision, Calibration Is Not Validation on why a small historical residual is not fitness for the decision, Dashboard Is Not Decision on why seeing a display is not a Decision Case and not a decision, Coverage Is Not Control on why the watched set is not permission to order work, Action Is Not Execution on why a recorded ACTION disposition is not the write, Recommend Is Not Authorize on why a drafted proposal is not permission, Verification Is Not Optional on why the case stays open until the check is recorded, and Silence Is Not Clearance on why a quiet dashboard is not authorization. A Reliability Assessment asks whether the records can support a conclusion. A Strategic Pilot is a governed proof around one operating decision. The Action chapter records intent. Authorized execution systems perform the write.