Mapping UML Techniques to Design Activities

Abstract

INTRODUCTION Over the years, the information system design process (Gero and Kazakov, 1996; Goldschmidt, 1997; Guindon, 1990; Jeffries et al., 1981; Parnas and Clements, 1986) has been investigated using a variety of perspectives. Researchers have examined cognitive aspects of design (Goldschmidt, 1997; Guindon, 1990; Guindon, Krasner, and Curtis, 1986; Rowe, 1987; Sen, 1997), design strategies (Adelson and Soloway, 1988; Batra and Antony, 1994; Guimaraes, 1985; Jeffries et al., 1981), and reuse tasks (Sen, 1997). A variety of modeling techniques, such as the entityrelationship model (Chen, 1976), data flow diagrams (Gane and Sarson, 1979), and object-oriented models (Booch, 1994) have also been developed to document the artifacts generated during the design process. Increasingly, the object-oriented design paradigm and related modeling techniques have been the choice of system designers. It is reasonable to expect that these modeling techniques (proposed to document the design products) will assist or at least not hinder the designer behaviors (that is, the process of IS artifact design). The expectation has, however, not been subjected to investigation. In this contribution, we focus on behaviors that designers exhibit during actual design sessions and investigate how these behaviors are supported by the de facto standard for object-oriented modeling, the Unified Modeling Language (UML) (Booch, Jacobson, and Rumbaugh, 1999). We focus on three core UML techniques: use cases, class diagrams, and sequence diagrams. To observe behaviors employed