Editing Rounding (section)

===Exact computation with rounded arithmetic===
It is possible to use rounded arithmetic to evaluate the exact value of a function with integer domain and range. For example, if an integer {{mvar|n}} is known to be a perfect square, its square root can be computed by converting {{mvar|n}} to a floating-point value {{mvar|z}}, computing the approximate square root {{mvar|x}} of {{mvar|z}} with floating point, and then rounding {{mvar|x}} to the nearest integer {{mvar|y}}. If {{mvar|n}} is not too big, the floating-point round-off error in {{mvar|x}} will be less than 0.5, so the rounded value {{mvar|y}} will be the exact square root of {{mvar|n}}. This is essentially why [[slide rule]]s could be used for exact arithmetic.