2 The data model

The data model is the heart of any running Ferret. From an abstract point of view, it is the whole object being edited by the tool. According with the general edition model, the Ferret application is the editor, and the data model is the object under edition. The user can then manipulate several aspects of the datamodel by using the appropiate view:

• Entities, relationships and class/subclass relationships are manipulated via the E/R editor, editing the conceptual view of the datamodel.
• Relations are manipulated via the relational editor, editing the logical view of the datamodel.
• Physical issues like cluster indexes and DBMS dependant data types are manipulated via the SQL editor, editing the physical view of the datamodel.

Figure 2.1: Data model

So the datamodel is an abstract object from which several views are made, at convenience. Note that each view (conceptual, logic and physical) has its own abstraction level, from higher to lesser. Because that the datamodel should maintain dispar information pertaining to different abstraction levels, such as entities, relationships, relations, cluster indexes, etc.

When the user changes some aspect of the datamodel (editing a view) the datamodel should reflect these changes in order to maintain the global coherence. So if the user adds a new entity to the datamodel, the corresponding table and SQL sentence should be also added. In this respect the datamodel views should be seen by the editors as opaque objects.

• Data model implementation
• Data model objects
• Object types
• The data model API