Databases Reference
In-Depth Information
models does not answer our need to initiate the construction
of one's own semantic model. These models are burdensome
to a company and the effort that is made to agree to take
them on board is considerable. Furthermore, they generally
only cover one portion of the necessary data and their
extensibility is not always evident;
-
the data model is expressed is technical terms, in the
form of an XML schema, or possibly a relational type, old
generation DDL. Business aspect modeling in UML does not
exist, or has been disintegrated. It is then difficult to
understand all the richness of the handled data;
-
the data model lacks a global architecture. It is rare
to benefit from a normalized model in the form of
components such as those which we have previously
described, i.e. business object domains, data categories and
business objects. There is too often a class imbroglio, with no
imprint, despite this being necessary, for a Enterprise Data
Architecture across the whole of the Information System.
8.5.3.
Generic data models
These models are less ambitious than those specialized by
industry. They have the advantage of being taken care of
rapidly and usefully contribute to the construction of its own
semantic model. These models are often more respective of
an Enterprise Architecture approach, with a real concern of
reusable semantic patterns.
These patterns highlight the boundaries that isolate
groups of data enabling a construction by stages of its data
model, following a LEGO approach. These models bring an
interesting approach to counter the blank page effect of the
first semantic model.
