AbstractTesting a program designed to recognize several dozen kinds of anomalies in large volumes of input data requires test data with many combinations of the anomalies in the data stream. The large number of combinations of anomalies makes it impractical to generate test data manually. This paper describes a test data synthesizer designed to help verify the accuracy, capacity and correctness of such a program. The operation of the synthesizer is analogous to the Fourier synthesis of a periodic function. A sweep action permits test data with well known response to be combined in a methodical way to explore complicated interactions which are not well understood.
展开▼