I maybe wrong but the current bounds commands (boundingBox and boundingBoxReal) come in rather large and look like they incorporate all vertices including things like memoryPoints for getIn position and the like.

Would it be possible to supply a bounds command that would only query vertices from a specified LoD like Geometry, so we can get bounds that are more relevant to the area actually occupied by the the visual model?

Currently doing boundingBoxReal on something like a "O_Truck_02_transport_F" produces a bound area that extends >4m front and back of the vehicle and about 2m above and below. I realise the front and below expansion is due to the largest offset from [0,0,0] but with their current returned values these commands have wasted potential for use as object placement mechanics.