The Data Modeler supports both top-down and bottom-up modeling.
To begin with, a complete model is created for the area to be modeled. As modeling proceeds, this model is further refined through the addition of entity types and submodels, which can, in turn, be further refined.
Attributes are entered for the various entity types. If the technical and semantic characteristics of the attributes have been fully clarified, data elements (and domains) representing these characteristics can be created or existing ones can be accessed from within the data model.
The link between the Data Modeler and the ABAP/4 Dictionary is established by assigning entity types to tables or views. You can create new tables or views from within the entity type. The attributes of the entity type can be included in the table as fields and the data elements that are created can be used elsewhere.
Once assignment to a table or view has taken place, the fields of the table or view are displayed as attributes of the entity type and are laid over the attributes (of the same name) previously created for the entity type. Changes to the assigned Dictionary objects (that is, creation/deletion/modification of fields) thus become immediately effective and visible in the Data Modeler too. Since both data model and Dictionary access the same information, their consistency is ensured.
The Data Modeler can also be used to remodel an existing application.
Entity types are created for the existing tables and views to which the tables and views are then assigned. By these means, the entity types obtain their attributes.
Logically connected entity types are combined to form data models. These can, in turn, be combined with other entity types and data models to form generic data models.
You can, of course, use a mixture of these two strategies for modeling and implementation of an application.