Of the articles in this repository, Randomization and Sampling Methods and More Random Sampling Methods combined are very long (about 230 KB in size combined). I would like to reduce their size while maintaining the most relevant algorithms for random variate generation.
Here are my goals for both articles:
- To shorten the Randomization with Real Numbers section as much as possible, while still describing the most general (and exact) algorithms possible for sampling real numbers of any distribution.
- To put emphasis on algorithms that work with random integers (or, if necessary, rational numbers), rather than random floating-point numbers.
- To put emphasis on algorithms that sample a distribution exactly, or at least with a controlled upper bound on the error. For discussion, see "Exact, Error-Bounded, and Approximate Algorithms".
- To ensure the documents are easy for programmers to understand and implement.
Letting them know: @dhardy, @lcrocker, @dj-on-github, @maciej-bendkowski, @christoph-conrads . If you know anyone familiar with random variate generation, let them know about this issue.