Testability, testing, and critical software assessment

Abstract

Although the phrases "critical system" and "critical software" encompass different degrees of "criticality" based on the user and application, I consider critical software to be that which performs a task whose success is necessary to avoid a loss of property or life. Software testability is a software characteristic that refers to the ease with which some formal or informal testing criteria can be satisfied. There are varying metrics that can be applied to this measurement. Software validation generally refers to the process of showing that software is computing an expected function. Software testing is able to judge the quality of the code produced. Software testability, on the other hand, is not able to do so, because it has no information concerning whether the code is producing correct or incorrect results. It is only able to predict the likelihood of incorrect results occurring if a fault or faults exist in the code. Software testability is a validation technique, but in a different definition of the term /spl Gt/OPEN validation" that the IEEE Standard Glossary of Software Engineering Terminology allows for. Software testability is assessing behavioral characteristics that are not related to whether the code is producing correct output.<>