The following algorithm finds the square root of a positive number:

 Algorithm squareRoot(number, lowGuess, highGuess, tolerance) newGuess = (lowGuess + highGuess) / 2 if ((highGuess - newGuess) / newGuess < tolerance) return newGuess else if (newGuess * newGuess > number) return squareRoot(number, lowGuess, newGuess, tolerance) else if (newGuess * newGuess < number) return squareRoot(number, newGuess, highGuess, tolerance) else return newGuess 

To begin the computation, you need a value lowGuess less than the square root of the number and a value highGuess that is larger. You can use zero as lowGuess and the number itself as highGuess. The parameter tolerance controls the precision of the result independently of the magnitude of number. For example, computing the square root of 250 with tolerance equal to 0.00005 results in 15.81. This result has four digits of accuracy.

Implement this algorithm.