What is Scientific Code

The aim of this chapter is to provide guidance on the challenges and approaches to testing computational applications. Testing in our case is focused on code testing for accuracy as opposed to validating the science models or testing user interfaces. A testing framework is used to present the different challenges. Discussions cover topics such as test oracles and the tolerance problem, testing to address specific goals rather than testing as a process, areas of risk inherent in developing and using computational software, a testing mindset, and the use of technical reviews. Three observational studies are included to illustrate different techniques, problems, and approaches. There is no prescribed way of testing computational code. Instead, an awareness of risks and challenges inherent in computational software can provide the necessary guidance.