Refinement and Extension of SMDM, a Method for Defining Valid Measures
Luis Reynoso (University of Comahue, Argentina)
Marcela Genero (University of Castilla-La Mancha, Spain)
Mario Piattini (University of Castilla-La Mancha, Spain)
Abstract: Although literature contains a huge amount of measures for measuring quality characteristics of software artifacts throughout the development life-cycle, the majority go no further than the step of definition. The key to obtaining valid measures which may be useful in practice is to carry out a definition of these by following a rigorous method. In a previous work we defined a method for obtaining valid measures, called SMDM (Software Measure Definition Method). In this paper we present the extensions and refinements of this method, which has been redefined in the light of seven years of application to various software artifacts, such as OCL expressions, UML diagrams, ER diagrams, Relational database schemas, Datawarehouse conceptual models, etc. In order to illustrate the redefined method, an example of the definition of a measure for the import-coupling of OCL expressions is presented.
Keywords: empirical validation, measure definition, psychological explanation, software measures, theoretical validation