Mälardalen University and Ericsson have found new ways to achieve productivity gains with software test

On Test Design is based on ten studies at a number of companies that aim for high quality in their software. In this thesis, several approaches to evaluating different test design techniques are presented – approaches that promote understanding of how causes for errors should be classified so that techniques can be compared between different systems in a general way.

– Our aim is to simplify testing and, in this way, to offer a better understanding of the processes surrounding test case construction and test design. By following our recommended guidelines for testing, product quality can be increased and development time be reduced. The guidelines may be of use in connection with studies in software engineering/programming, too, in that they help to indentify the kind of knowledge about tests that is important to convey, says Sigrid Eldh.

Testing is the dominant method for quality control of industrial software today. The approach implies that a system is tested in a controlled environment in order to discover errors and deviations from the expected behaviour. Although companies put a lot of money into testing, they are in great need of better test design techniques, the area is surprisingly underdeveloped, which makes this research project one of a kind.
– We have studied why advanced test design techniques are not used when test cases are constructed in companies, and we have found causes and pitfalls related to the methods that explain why many systems are so poorly tested. We have developed concrete guidelines for the testing of industrial systems that, thanks to increased efficiency and enhanced product quality, can create great value for society, Sigrid Eldh rounds off.