Scientific Software Testing: A Practical Example

Abstract

When developing scientific applications, scientists are mostly interested in getting scientific research achievements. Usually, the first phase in scientific applications development process is coding. Testing is rarely performed or it is not performed if the final result is correct. Many bugs are found later and problems arise when no software documentation can be found. The goal of this paper is to improve the testing of scientific applications by describing the full testing process which follows the software engineering testing principles. In order to confirm the usefulness of the process, we develop and test an application for solving 1D and 2D Schrödinger equation by using the Discrete Variable Representation method (DVR). Using this testing process in practice resulted in requirements specification, test specification and design, finding connection between specified requirements and tests, automated testing and executable tests.