Test-Management

A major challenge in developing standard software is achieving a certain quality of the developed product. Therefore, the ready developed product is tested. In doing so, one tries to reach a preferably high test coverage of the product's functionality with individual tests. Organizing such tests, the resulting correction of bugs, and testing a specific functionality anew, is a major challenge in today's high quality software development.

Experience has shown, that only 20 percent of the test cases can detect 80 percent of a software product's bugs. As a general rule, well-known methods for test design and determination of test end criteria allow for designing exactly those test cases comprising the decisive 20 percent of the test cases. The ever increasing problem in today's software development process is rather an economic one: Temporal restrictions, narrow budget, pressure of competition, and short time to market often do not allow for conducting all 100 percent of the scheduled tests.

For the system test, smaller and middle-sized projects often require hundred and more tests, bigger projects often several thousands. To trace the development process as promptly as possible and to verify correction of faults detected in previous passes, all these tests should be conducted, yet several times and in parallel.

Soft-net Austria addresses the abovementioned issues in terms of developing a research prototype (the so called TestCockpit). The TestCockpit is intented to serve as a central tool for the test coordinator of saftey-critical, highly-availabe or business-cricitcal software. Even in the context of large scale projects the TestCockpit allows for monitoring and tracking the development process, the achieved test coverage, and the state of fault correction. A central issue is on the overall view among all conducted test activities as well as on providing the evidence of conducting the tests successfully.