The detail allows handlers to distinguish between high volume cursor motion events and everything else. This allows (most) event handlers who don't care about cursor motion to not have their power/CPU usage impacted by it.
Using minimal classification is important because it allows those handlers
who do want to use this optimization via
captured-event::nonmotion to keep
supporting multiple event types without requiring any logic changes.
Any handler that does want cursor motion events and others should simply
remain unmodified and keep using