HA Light Controller

Purpose

Home automation platforms generally treat a command as a result. Send “on”, record “on”, move on. Radio-controlled lights do not cooperate with that assumption, and the gap between the recorded state and the room is where the frustration lives.

This integration is the smallest honest version of the fix: command, then look, then act on what was actually observed.

Operating model

A command is followed by a read of the achieved state. If the two disagree, the integration retries within a bounded policy rather than reporting success or failing silently. State that could not be confirmed is represented as unconfirmed, not as the state that was requested.

The bound matters as much as the retry. An unbounded retry loop against an unresponsive device is a new failure, not a fix for the old one.

Practices

Observe is the read-back. The system’s picture of a light comes from the light.

Verify is the comparison between commanded and achieved state, which is the entire premise.

Recover is the bounded retry and the honest unconfirmed state — a path back to a correct picture without pretending the first attempt worked.

Architecture

The integration is small on purpose. It sits between Home Assistant and the device, adds a verification step to a command that previously had none, and does nothing else.

The rejected shortcut was a fixed delay before assuming success. A delay long enough to be reliable is long enough to be annoying, and it still guesses.

Evidence

The repository is public under MIT, with published releases suitable for installation into a Home Assistant instance.