To compute the relative error (assuming you're trying to find a stationary point of the function), after computing ff, subtract ff from x and take the absolute value. Then maybe you should divide that by the absolute value of x, but that depends on how you define relative error.
Then use a conditional statement (an "if" statement) to test whether the error is less than the tolerance, and use "break" to stop the loop if it is.
Other points to note:
- 0.00001% is 1e-7, not 1e-5.
- You can save computation time by using fx = ff instead of fx = eval(f). You do this after testing the error of course.
