Re: [math-fun] Origin of SQRT hack?

I'm sort of coming into the middle of this discussion, but I think that shifting to get the square root of a floating point binary number must be relatively old. I vaguely recall mention of the strategy for the ORDVAC at Aberdeen Proving Ground around 1954, but a more verifiable reference might be to Konrad Zuse's machines of the 1930's. As I recall his biography mentions that a fast and reliable square root operation was one of the notable features of those machines. If shifting weren't as expensive as multiplication, you could probably get fast squares the same way. -hvm