How many terms in the truncated continued fraction does it take to approx- imate ? with an error less than 10 - 10 ? As the number of terms increases beyond this roundo?, error eventually intervenes. What is the best accuracy you can hope to achieve with double-precision ?oating-point arithmetic and how many terms does it take?