Invent before Saturday

Propose a one-wall pilot for Kiln Hall. The goal is not to model the whole town. The goal is to give Mara enough timely, private information to decide whether the north wall needs attention before Saturday class.

Write a proposal with these parts:

• Scope: name the wall, nearby pipe section, and the decision the pilot should support.
• Measurements: state what will be measured, where each sensor will go, and why that reading could help identify an early problem.
• Privacy: state explicitly that the system collects building conditions, not video, audio, names, or attendance.
• Response: say how readings reach Mara, what counts as an alert, and what she does after a false alarm or a concerning reading.
• Maintenance: name who checks the sensors, updates the model after repairs, and reviews whether the pilot is useful after one month.

Keep the test concrete. For example: “Install one flow sensor on the north-wall supply line and two moisture sensors at known risk points. Send an alert to Mara if flow rises outside the hall’s normal closed-hours range. Mara inspects the area or calls the maintenance contact; no cameras or microphones are installed.”

A good proposal includes a way to learn from errors. It does not promise that a sensor will predict every failure.

Common mistake: proposing a large platform before defining one decision it must improve.