That, as they say, is a very good question. I'm intrigued as to what you have in mind. Maybe you have a feature for the ? ? ? tab.
The answers: Currently my 2.5D probing routine will not cope with not being able to go all around the object. Why? because it wasn't designed to do that and it will hit limits, soft hard or physical.
Re: the efficiency of my 3D routine: The less rectangular the object, the more efficient it is over a "standard" bed o nails routine. I'm thinking there's probably a nice mathematical equation for calculating the actual efficiency and it might be based on something like this: Calculate the area of the rectangular bounding box of the object. Calculate the actual area of the object. Divide the former by the latter giving a measure of efficiency. Obviously the more real estate that is NOT probed by mine but that IS probed by the standard, then the more efficient mine is.
Hope this helps.
Ian