• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

System validation turns out to be a bottleneck to the technological progress, as the inability to adequately evaluate systems may become the limiting factor in the ability to employ systems that the progress of science and technology will allow to design. Therefore, it is the aim of the present paper to provide some formally well-based contribution to the area of system validation. In contrast to system verification which deals with the problem "to build the system right", we adopt the view that system validation deals with the problem of "building the right system". Consequently, the task of system validation is the conceptually more difficult one, as its success has to be measured against the usually only incompletely and vaguely given expectations of some user or, even worse, of some dynamically changing community of users. Theoretically, a simple way to prove or disprove crucial properties of some target system is to investigate all its input/output sequences. But the exponential growth of possible input values in dependence on the input quantity sets narrow limits to an exhaustive checking of all input/output data. Motivated from the theoretical concept of a functionally exhaustive set of test cases and driven by the necessity to minimize test data drastically, we are led to the concept of a quasi--exhaustive set of test cases which is carefully introduced and formalized. There will be developed some fundamental scenario of appropriately using quasi--exhaustive sets of test cases for system validation.

Citations