Testability is a quality factor used to predict the amount of effort required for software testing and to indicate the difficulty of revealing faults. Also, it can be used to estimate the difficulty of software testing. This paper presents a quantitative testability analysis method for a software component that can be used when the source program is not available, but the bytecode is (as in Java .class files). This process analyzes the testability of each location to evaluate the component testability. The testability of a location is analyzed by computing the probability that the location will be executed and, if the location contains a fault, the execution will cause the fault to be revealed as a failure. This analysis process helps developers measure component testability and determine whether the component testability should be increased before the component is reused. In addition, low testability locations are identified. This paper uses an example to evaluate the ability of this process to indicate component testability.
Back to my home page.