Harry Sneed

Date: April 17, 2018
Time: 10.00 – 17.00

Requirement-based Testing



Next to providing a base line for developers, the requirement specification also serves as a reference for the developers. Testers need an Oracle to which they can refer. This was recognized and published by William Howden in the very first large scale testing project for the American Ballistic Missile Defense System back in 1977. Every test is a test against something. This something can be a model or another program. Here it is the requirement document. The actual behavior of the system under test is being compared with the behavior prescribed in the requirement document. The prescribed behavior is captured in the requirement text, which means that text has to be interpreted, independently of the programming language it is implemented in. Thus, the requirement text is aimed primarily at the testers. They should be able to convert that text into unique test cases with pre- and post-conditions, and representative test data. Just as automated unit test cases are extracted from the code, automated system test cases are extracted from the requirement specification or model.

Today after 40 years, the testing community returns to that original testing approach. One speaks now of model-based testing, but requirement documents are seldom models. They are texts and the approach taken is to parse the text to extract the test conditions, i.e.paths, from it. In this half day tutorial participants are made familiar with the principles of requirement-based testing and shown how it is done using sample requirements from German industry.



1) Introduction to Requirement Specification – RXVP & RSL
2) Requirement vs. Model-based Testing
3) The significance of Use Cases in defining Test Cases
4) Test conform Structuring of the Requirements Document
5) Examples from industrial Practice