This paper contains a brief description of the R2M (Role and Request Modeling) method and its supporting visual modeling CASE (Computer Assisted Software Engineering) tool. R2M is a modeling method for creating Conceptual Models of work systems using a combination of ontological and object-oriented concepts. Ontological principles serve to define the meaning of modeling constructs in terms of domain semantics, and to derive rules guiding the modeling process. The CASE tool is a graphical software tool that supports the creation of models according to the R2M method. Guided by the principles of R2M, the tool helps assure the semantic integrity of models, and enables management of complex models via decomposition (i.e. more details at decreasing abstraction levels). The tool can help ensure consistency between different modelers and completeness of models.
actors that assume certain roles. This view led us to develop a set of modeling rules which address two issues: first - the mapping of domain phenomena to a model; and second - semantic integrity constraints that can be applied to constructed models. Based on these rules, we developed a modeling procedure that assures the ontological validity of constructed models. The procedure can identify situations where the modeler needs to clarify domain aspects with stakeholders.
The modeling approach – termed Role and Request Modeling (R2M) – has been implemented in a CASE Tool. This tool embeds data structures that reflect the fundamental ontological concepts and principles (that in turn guide the semantic integrity rules). As well, the tool provides checks for the adherence of constructed models to the modeling rules.
R2M is graphic notation-independent. However, the user interface of the R2M software (shown in Figure 1 below) uses an intuitive representation of the modeled domain. The information about the model appears in several visible panes: • The Role Explorer (left side) displays all roles in the model for easy navigation. • The Modeling Canvas (main portion) in which the model is created by the user. • The Property Details (lower portion) where details about the role currently selected in the Modeling Canvas are displayed and manipulated. • An additional pane showing errors in the model (the Semantic Errors pane, described below) can be visible or hidden.
To enable construction of models of a complex environment R2M supports a wellformalized method of and rules for decomposition. The rules assure that models at any level will be ontologically and syntactically consistent with higher and lower level models of the same domain.
As an example of decomposition using R2M, Figure 2 shows part of a domain model within the Modeling Canvas. The view shown is the top level model – i.e. the highest level of abstraction. At this level in the example, both the “Customer” and the “Office Clerk” roles communicate with the “Warehouse” role.
Figure 3 shows part of the decomposition model of the “Warehouse” role of the same domain. This view shows roles and communications that are internal to the “Warehouse” (i.e. the “Warehouse Manager” and “Warehouse Worker” roles, and the communications between them) as well as the communications between these and the roles that – at the top level – appear to communicate with the “Warehouse” role. As can be seen in this example, the “Office Clerk” communicates with “Warehouse Manager” and the “Customer” communicates with the “Warehouse Worker”. R2M supports decomposition to any level and ensures consistency between the levels.
We have experimented with the R2M method and tool both in teaching situations and in practical (and realistic size) cases. The results have shown that the use of the method led to consistency of models across modelers. Furthermore, semantic errors identified by the tool were often an indication for the analysts to seek additional information about the modeled domain, thus leading to more complete and accurate models.