3.2 Steps for optimizing the object model

The observations made at the beginning of this section imply the following rules for optimizing the object model:

• Identical attributes and operations appearing in several different object classes are gathered in a (eventually new) superclass (see 3.2.1).
• Attributes defining the type of different instances of an object class are removed and replaced by appropriate subclasses (see 3.2.2).
• ''pushbutton'' variables become methods of the corresponding class (see 3.2.3).
• Relationships between object classes should be as accurate as possible; containment and inheritance are introduced into the model (see 3.2.4).
• New data types (e.g. enumerations) are defined for attributes with restricted value ranges (see 3.2.5).

The results of the application of these rules are described in the following subsections.