Abstract

When applied to real-world problems, object-oriented modeling maps an entity in the real world to a class as it is. This seemingly natural “genuine” real-world modeling can be rightly applied to cases when the purpose of modeling is merely to represent a problem in a class diagram and thus to facilitate its understanding. Business process reengineering is a good example of this. Genuine real-world modeling can also be applied to the development of a program that simulates the real world on a computer. Contrary to these cases, however, “pseudo” real-world modeling has instead to be applied when a business assistance application is to be developed. It maps an entity whose information is dealt with by the business to be automated to a class that represents the information about the entity. These two modeling methods have to be appropriately applied according to the type of their target problems. This point, however, has not been sufficiently recognized. Many authors of the literature on object-oriented methodologies and techniques teach us “naive” real-world modeling, whose real nature is a mixture of genuine and pseudo real-world modeling methods. Naive analyzers who believe the literature are lured into severe modeling errors when they develop business assistance applications.

Author Keywords: Object-oriented methodology; Object-oriented analysis; Real-world modeling; OOSE; OMT; UML

Article Outline

1. Introduction

2. Two object-oriented modeling methods

3. Rules of pseudo real-world modeling

3.1. Scope of modeling

3.2. Operations assigned to classes

3.3. Definition

3.4. Applying pseudo real-world modeling to examples

3.4.1. Example 1

3.4.2. Example 2

4. Errors due to naive modeling

4.1. Example 1

4.1.1. The scope of modeling

4.1.2. Operations assigned to classes

4.1.3. Three categories of errors

4.2. Example 2

5. Discussion

5.1. Modeling a cyber world

5.2. Modeling the system together with its environment

6. Conclusion

References

Vitae