Abstract:The paper presents results of analysing and evaluating the OPEN Modelling Language (OML) metamodel in terms of the Bunge-Wand-Weber (BWW) model of information systems and the Framework of Information Systems Concepts (FRISCO). The analysis and evaluation has suggested several ways to improve the modeJling constructs supported by OML and to improve how OML is defined and organised. It has also made it clearer how and why ontological models like the BWW -model and FRISCO can and should be used to improve and define object-oriented (00) modelling languages. In particular, it has led to a template for defining 00 modelling constructs in terms of what they represent in the problem domain during IS development. Finally, the analysis and evaluation has suggested ways to improve FRISCO as an analysis and evaluation tool for 00 modelling languages.
INTRODUCTIONObject-oriented (00) modelling languages have become widespread in recent years. However, some critics claim that object orientation is less well suited for modelling "real-world" problem domains during information systems (IS) development because object-oriented abstractions are tailored to represent software artefacts rather than things in or aspects of problem domains. For example, critics question whether OO-concepts such as message passing and encapsulation are really features of the real-world Opdahl, Henderson-Sellers and Barbier things that are represented and whether things really belong to fixed classes throughout their life-times (e.g., Jackson, 1995; Parsons and Wand, 1997b.) To shed light on these problems and related ones, the paper presents results of using ontology to analyse and evaluate how well the metamodel of the OPEN! Modelling Language (OML; supports representation of "real-world" problem domains. The analysis and evaluation is anchored in the Bunge-Wand-Weber (BWW) model of information systems (e.g., Wand and Weber, 1995;Weber, 1997) and in the Framework of Information Systems Concepts (FRISCO;Falkenberg et al., 1998.) It is the result of systematically comparing more than 160 OML-constructs with 44 ontological constructs in the BWW-model and 98 constructs in FRISCO.Since space does not permit a detailed discussion of each OMLconstruct, the paper instead discusses important outcomes in general.(1) The analysis and evaluation has suggested several ways to improve the modelling constructs supported by OML and to improve how OML is defined and organised. (2) It has made it clearer how and why ontological models like the BWW -model and FRISCO can and should be used to improve and define object-oriented (00) modelling languages. (3) As a consequence, it has led to a template for defining 00 modelling constructs in terms of what they represent in the problem domain of IS development. (4) Finally, the analysis and evaluation has suggested ways to improve FRISCO as an analysis and evaluation tool for 00 modelling languages. Since detailed suggestions for improving OML will be presented elsewhere, the paper will focus on the latter th...