Empirical Verification of TQED - A New Test Design Heuristic Technique

TQED is a universal test heuristic that assists testers in creatively designing effective test cases. It involves defining the test problem in terms of component elements, each of which is classified into one of the four so-called dimensions, which are: time (T), quantity (Q), event (E) and data (D). Then, test ideas are created by considering specific combinations of the components, aided by the interpretation of combinations of dimensions. In this article, we compare the TQED model with other well-known test heuristics and risk analysis techniques, and then present an empirical verification of the effectiveness of the TQED model. We compare the effectiveness of tests written by 24 developers who were asked to implement code for the same problem, together with the unit tests. The developers were divided into two groups, one of which used TQED when designing unit tests and the other did not. Effectiveness was measured in terms of code coverage, mutation coverage and failure rate of test cases. To increase the objectivity of the study, a cross-experiment was conducted in which each developer's tests were run on the source code of all other developers. Our research showed that TQED can significantly support testers in creating strong tests that are more likely to detect defects in code.