Conceptualizing systems for understanding: An empirical test of decomposition principles in object-oriented analysis

During the early phase of systems development, systems analysts often conceptualize the domain under study and represent it in one or more conceptual models. One of the most important, yet elusive roles of conceptual models is to increase analysts' understanding of a domain. In this paper, we evaluate the ability of the good decomposition model (GDM) to explain the degree to which conceptual models communicate meaning about a domain to analysts. We address the question, ‘Do unified modeling language (UML) analysis diagrams that manifest better decompositions increase analysts' understanding of a domain?’ GDM defines five conditions (minimality, determinism, losslessness, weak coupling, and strong cohesion) deemed necessary to decompose a domain in such a way that the resulting model communicates meaning about the domain effectively. In our evaluation, we operationalized each of these conditions in a set of UML diagrams and tested participants' understanding of those diagrams. Our results lend support to GDM across measures of actual understanding. However, the impact on participants' perceptions of their understanding was equivocal.