> To reciprocate a unum, you ignore the first bit and negate what remains! > Notice that exact numbers have bit strings ending in zero and open intervals betwe

> To reciprocate a unum, you ignore the first bit and negate what remains! > Notice that exact numbers have bit strings ending in zero and open intervals between exact numbers have bit strings ending in one. So reciprocating an inexact interval gives an exact number, and vice versa? > If you ask for 2 + 2 at this very low precision, there is no exact representation for four, so you have to use the open interval (2, 5), for example. So what is 2 + 2 - 2? I feel like this guy keeps contradicting himself, e.g. switching between impractically high and low precision depending on the question.