K.: Systematic vs. Operational Testing: The Necessity for Different Failure Models, to appear in

Software Reliability Growth Models assume that the application under test (AUT) is tested according to its operational profile. However, examining the test practices at different industrial test laboratories this is often not the case. Instead, a systematic approach to testing is taken. The differences in systematic versus operational testing, in particular their effect on the failure count model, is the topic of this paper. After a detailed discussion of the advantages and disadvantages of systematic and operational testing, the hypergeometric failure model by Rivers is presented. In our work problems in the original formulation of the model are resolved and the derivation of the continuous form is more stringent. A different approach to the interpretation of testing efficiency is taken. The model is applied to data collected during a test project at imbus AG and the results of the failure prediction are presented. The paper concludes with an outlook on further research activities. 1 Systematic vs. operational testing One possible way of testing software is using it according to its operational profile. This means that the operations- major system tasks of short duration with processing substantially different from other operations [9]- of the application under test (AUT) have to be determined together with their execution probabilities. Based on these probabilities the number of test cases is allocated to the different program functions. Th