Exercise 1.46: Several of the numerical methods
described in this chapter are instances of an extremely general computational
strategy known as 
iterative improvement.  Iterative improvement says
that, to compute something, we start with an initial guess for the answer, test
if the guess is good enough, and otherwise improve the guess and continue the
process using the improved guess as the new guess.  Write a procedure
iterative-improve that takes two procedures as arguments: a method for
telling whether a guess is good enough and a method for improving a guess.
Iterative-improve should return as its value a procedure that takes a
guess as argument and keeps improving the guess until it is good enough.
Rewrite the sqrt procedure of 1.1.7 and the
fixed-point procedure of 1.3.3 in terms of
iterative-improve.