Postby Technomancer » Sun Aug 07, 2005 3:29 pm
It depends on what you're doing, and what you plan to be working with, and what kind of probability distribution you need. The C-function rand for example works fine, but only gives you a uniform distribution. GNU Scientific Library (more C) also has a number of different random number generators, which seem to work well enough and also covers a number of different distributions (e.g. Gaussian, Cauchy, etc, etc). I believe the book 'Numerical Recipes in C' also covers some of the programming aspects.
The basic RNG's in MATLAB only cover uniform, Gaussian and Poisson distributions, but again you should be able to program other pdfs with a little bit of work. MATLAB is easy to use, but personally I'm not very satisfied with their RNG, since it outputs the same results in sequence each time you load the program.
Another alternative is to simply look in books of scientific/engineering tables, which frequently have tables of random numbers generated according to some distributions. While they can be convenient at times, these books are also fairly hard to find these days owing to the appearance of digital computers.
The scientific method," Thomas Henry Huxley once wrote, "is nothing but the normal working of the human mind." That is to say, when the mind is working; that is to say further, when it is engaged in corrrecting its mistakes. Taking this point of view, we may conclude that science is not physics, biology, or chemistry—is not even a "subject"—but a moral imperative drawn from a larger narrative whose purpose is to give perspective, balance, and humility to learning.
Neil Postman
(The End of Education)
Anti-intellectualism has been a constant thread winding its way through our political and cultural life, nurtured by the false notion that democracy means that my ignorance is just as good as your knowledge
Isaac Aasimov