Risks and Benefits

One of our goals was to characterize each representation strategy in terms of its attendant risks and benefits. We made only piecemeal progress along these lines, but believe that the framework as a whole could be annotated with these criteria. Some observed were as follows:

1. Using a single system model as a domain model:
   • Benefit: Use of existing notation, technology; no need for additional training, tools etc.

   • Risk/Limitation: Can only model a single system and its specific implementation. Contains all system-specific decisions, constraints, etc. Does not capture multiple systems perspective.

2. Using OO notation:
   • Benefit: For an organization that has already made a commitment to OO techniques, use of OO notation builds upon already-familiar notations.

     In addition, use of a common notation allows for tight integration of the domain and system models.

   • Risk: Confusion with respect to terminology; overloading of semantic elements.

   • Mitigation: Adopt conventions that clarify which portion of an OO model is being used as a domain model vs. a single-system model. (The use of frameworks seems to be one approach along these lines.)

We would hope to expand these criteria across the full set of strategies identified.