The current definition of Device.property is:

Characteristics or features of the device that are otherwise not captured in available attributes, e.g., actual configuration settings, time or timing attributes, resolution, accuracy, and physical attributes. The focus is on properties of the device actually in use while DeviceDefinition focuses on properties that are available to be used.

The R4 definition is simply the first sentence.

Previously I understood that property was intended to capture "fixed" properties (this is a "large" blood pressure cuff, this bed can support up to 900 lbs, the resolution of this MRI is up to [some number]). However, inclusion of "actual configuration settings" and "properties of the device actually in use" leads me to think that the intent is to record the value of dynamic settings in effect at any time. If that is the case, there is overlap and confusion about the role of DeviceMetric (does device metric define a property?), and of "settings" Observations, which refer to those metrics. In my mind, DeviceMetric records the existence of a setting (there is a knob to adjust the volume) and a "setting" Observation captures the value of that setting at a point in time (our volume is 11). However, property seems to be described as supporting the "settings" Observation behavior. If that is the case, clarify how to record a device setting at a point in time, and track changes to that setting over time.

The boundaries and use of these resources, and elements needs to be cleared up.