The control panel exists so any diagnosis can be produced again on demand.

The chain

Fourteen functions per block, and each one is either a reference or a rule.

A composite's weekly block is a fixed chain with no discretion anywhere in it. Cumulative efficiency and cumulative profit factor are referenced from the structural truth tables of their respective pipelines. Each is converted into a class against its own spectrum. The two classes concatenate into a matrix key, and that key retrieves a base state and a functional role from the lookup tables. The primary drift readings for both metrics are referenced from their interpretation layers, and their relationship yields the conflict type. Both conviction readings are referenced likewise, and severity is derived from them. The final diagnosis combines base state, conflict type and severity into one line. Every step is either a reference to a value computed elsewhere or a rule applied to two values already present — which is why the chain reproduces exactly.

Two selectors, one address

A week and a composite together name exactly one diagnosis.

The addressing scheme is what makes the engine's history usable rather than merely stored. Any statement the engine has made resolves to a pair — this composite, that week — and the pair is sufficient to regenerate the statement in full. This matters most when a past decision is being reviewed. An operator revisiting a call made three months ago does not need to remember what the engine said, and does not need to trust a note written at the time; the selectors go back and the diagnosis reappears, including the base state, the conflict type, the severity and the conviction readings that produced them. That property is what separates a diagnostic history from a log. A log records what was concluded. This regenerates why.

The status fields

Mode, data source and engine status say whether the reading can be trusted at all.

Three of the control panel's six fields are not selectors but declarations about the engine's own condition, and they carry the same discipline the rest of the system applies to uncertainty. The data source states where the values are coming from. The mode states how the engine is being run. The engine status states whether it is functioning. Their presence means the surface has three honest states rather than two: the reading is valid, the engine is not currently configured or sourced properly, or something has failed — and the third is announced rather than expressed as a plausible-looking diagnosis built on stale inputs. Reading these before the diagnosis is the same discipline as confirming the cockpit's intake values, and it exists for the same reason: a translation surface will translate whatever it is handed.

The latest-week field

Showing the latest available week beside the selected one prevents a silent staleness.

The panel displays the latest available week alongside whichever week is currently selected, and the pairing does something specific. A selector left on a past week will render a complete, correct, fully-formed diagnosis for that past week, and nothing about its appearance indicates it is not current. This is the same failure as a stale routing value in the cockpit and it is caught the same way — by putting the two numbers in the same field of view, so the mismatch is visible in the glance that reads the diagnosis rather than requiring a separate act of remembering. It is a small piece of interface design carrying a real protection, and it follows the standing principle that a safeguard depending on the operator recalling something at the moment of temptation is not a safeguard.

Why it was built last

A control panel summarises an engine; it cannot be designed before the engine exists.

The note that section one was wired last is a build-order observation with a design lesson inside it. A control surface is a claim about what the machinery beneath it can be asked, and writing that claim first tends to produce an interface that constrains the engine to whatever seemed askable at the outset. Building the calculation grid, the compression strip and the counters first, then adding the panel that addresses them, means the selectors expose what the engine actually supports. It also means the panel is thin — six fields for an engine of considerable internal complexity — because most of what a reader needs is in the grid, and the panel's job is only to say which slice of it is on screen and whether that slice is sound.

• Any diagnosis is addressed by week plus composite, and regenerates from that pair.
• Mode, source and status give the surface a third state: broken, said out loud.
• Latest-week beside selected-week catches staleness in the same glance.

The key idea

Reproducibility is what turns a diagnostic history into evidence.

A record of past conclusions is worth something, but it cannot settle an argument about whether a conclusion was reasonable at the time, because the record contains the conclusion and not its basis. An engine whose every output regenerates from two selectors carries its basis with it permanently. That is what allows the system to be audited against its own past — to ask whether the engine has been calling divergences that later proved real, and to answer with the original readings rather than with recollection. The counters that track patterns across weeks depend entirely on this: they are only meaningful if each week they count is a week that can be produced again.