Main points:

  • standard software testing asks if we get the right answer to a particular problem
  • an alternative is to generate random problems and try to infer problems
  • this can more thoroughly test the space of problems
  • random input testing can be fun

annotated slides (pdf)

Presented 2011 August at useR!2011.

Going further

There is a bit more on this in Tao Te Programming.

Kommentarer inaktiverade för Random Input Testing using R

See more

Explore more content and blog posts.

  • jun 25, 26

    The email address patrick@burns-stat.com was out of action for a few hours today.  It is back now.

  • jun 25, 26

    Customization in R. Basics Several features benefit from being customizable — either because of personal taste or specifics of the environment. The way R implements this flexibility is through the [...]

  • jun 25, 26

    How to control the limits of data values in R plots. R has multiple graphics engines.  Here we will talk about the base graphics and the ggplot2 package. We’ll create [...]