Testability is a measure of how easy (less complex, less tedious, less boring, less costly) the effective testing of implementation is made.;In Object-Oriented development, testability is a result of (1) the characteristics of Object-Oriented implementations, (2) the test support environment, (3) the characteristics of representations (i.e. notations) and (4) the software process in which testing is conducted (67). In this thesis, we examine in details the last two factors.;When addressing the characteristics of notations, we determine that notations should exhibit the basic testability attributes of completeness, correctness, consistency and incorporation. In order for notations to exhibit these attributes, they must be verifiable, modifiable and traceable. The latter three are referred to as testability requirements.;Testable notations provide the ability to (1) uncover errors early in the process, (2) generate test cases, (3) guide the testing of the implementation and (4) facilitate the location of source(s) of error(s) for modification.;We identify testing activity types that are decoupled from the syntax of notations and from development methodologies. These testing activity types define what needs to be tested (e.g. different semantics) as well as coverage criteria. They are generic in the sense that they are refinable (i.e. adaptable to different levels of detail) in order to be uniformly applied at all phases of development.;When addressing the software process in which testing is conducted, we propose that there must be an integration of development and testing activity types. Also, we propose a way of handling change in a consistent manner such that testability attributes are not lost. Examples of these changes include error correction, accommodation of new requirements, etc. (Abstract shortened by UMI.)
展开▼
