ALT Most large digital computer facilities have available a short routine which will generate on call a random number (or more properly, a “pseudorandom” number) from the uniform distribution in the unit interval [ll, 121.We shall denote such a random number by r. By definition, the probability that a generated value r will fall inside any given subinterval of the unit interval is equal to the length of that subinterval, and is independent of its location: For o<a<p<1, Prob{cu < r < j?>= /3 - CL (22) We shall take it for granted here that we have ready accessto some such “uniform random number generator.” Our object now is to develop methods for using the output values r of a uniform random number generator to generate a random pair (T, r-l>according to the probability density function P(T, p) in (17).