The utility of point-in-time and interval events has become clear
during archetype design analysis of clinically measured phenomena.
To further improve the representation of interval events, we should
allow for:
optional record how many samples were taken in each interval event
(and/or sample width of original sampling leading to an interval
event).
variance and/or other measures during interval
a way of allowing one interval event to capture samples within +/-
xxx, i.e. outer limits to still call it one event.
to make archetyping clearer, maybe separate out EVENT, POINT_EVENT,
INTERVAL_EVENT