CppCon 2016

The generation of pseudo-random numbers has been an important computing application ever since the pioneering work of John von Neumann and others during the World War II era. It remains an important application for gaming, statistical sampling, and simulation, to name but a few areas of its continuing utility.

Obtaining a random number (technically, a variate) is widely believed to be a simple task. Alas, it usually isn't, and programmers often have little or no training to help them make important decisions that can have subtle effects on the correctness and reliability of their programs' results.

In this talk by the principal author of the C++ <random> facility, we will describe the header's conceptual underpinnings, demonstrate its most common correct usage pattern, exhibit and explain what's wrong with its most common usage anti-pattern, and discuss the merits of a number of possible toolkits based on the header's facilities.