Since the maintenance of the existing system and incremental development of new functions which have to be integrated are central tasks, a systematic testing method that has regression capability becomes more and more important. Until now, module and user interface testing is performed intuitively and there is neither knowledge about a systematic test method nor about tools supporting module and user interface testing.
The experiment may be characterised as follows:
- Education in systematic module testing
- In order to successfully introducing a systematic test method, the project members should have sufficient knowledge.
- Definition of a special test process for module and user interface testing.
- Based on the objectives we try to achieve with the systematic test method, we will define a small but efficient test process.
- Evaluation of tools supporting the systematic test approach
Although the introduction of a systematic test method is the central focus of the experiment, we should define a set of tools supporting the developers and especially the testers during the test activities:
- Introducing the tool in the development environment. In order to use the evaluated tools efficiently, we have to introduce the tools in the existing development environment.
- Performing tests according to the test process. Tests will performed according to the defined test process. This will include both module tests and tests of interactive user interfaces.
- Measuring aspects of the test process. In order to demonstrate the impact of the introduced test process on the overall software quality, we have to collect relevant data.
- Analysis of the experiment and integration of the test process. The last step of the experiment consists of the analysis of the experiences and of the integration of the defined test process in the overall software process.
EXPECTED IMPACT AND EXPERIENCE
The major commercial impact is increased competitiveness in non European markets. This shall be achieved by highly reliable software systems that are of good quality and therefore are attractive for the market. Only if the end-users, especially in the area of control system software, can trust in the effectively and reliability of their software systems, they can be convinced to invest in their software in the longer term.
To achieve the required software quality and reliability, the software test process employed has to become more systematic and more mature than it is today.