Cost Reduction through Combining Test Sequences with Input Data

Abstract

Nowadays computer applications mainly depend on input data that bear additional constraints to be considered and evaluated carefully. Very often, this invokes test sequences that describe the path to the place where input data is to be entered, e.g., while testing graphical user interactions. Model based testing allows deriving those test sequences from a model, for example, based on event sequence graphs (ESGs) where the nodes represent events. Unfortunately, combining ESG with the great variety of input data considerably inflates them with respect to the number of their nodes. To avoid this event inflation, previous work extended ESGs by decision tables to take this variety and constraints on input data into account. This paper extends the previous work and shows how to derive input data-supplemented test sequences and, at the same time, to considerably reduce effort for test generation and test execution. A case study drawn from a large commercial web portal evaluates the approach and analyzes its characteristics.