Karnaugh-Veitch Maps as Minimal Formal Contract Between Textual Requirements and Tests: A Use-case Based Technical Analysis

Abstract

Checking that requirements written in natural language hold for a formally implemented system is a complex task. Test steps are commonly implemented manually from the requirements. This process is inherently prone to mistakes as test cases are complex and need to be analyzed sequentially to check, which input/output combinations are tested (although tools allow for explicit tracing). Utilizing Karnaugh-Veitch maps as minimal formal contract between informal requirements and implemented test steps improves this process. KV-maps provide the requirements in a computer-editable way, as the correspond to Boolean formulas. KV-maps further allow to define which test steps are relevant. With both requirements and relevance specification at hand, test steps are automatically generated. The approach is applied on a real-world industrial use-case, a train control management system. Although being generally amenable to permutation testing, the selected use-case emphasizes the potential of the method. The method successfully demonstrates its benefits and may help to disclose flaws in the current manually implemented tests.