Databases Reference
In-Depth Information
the type of data which is handled to be defined: “simple data
type”, “complex data type” and “table”:
− “simple data type” is formed by a single piece of
information, without any nested data types. This is an
elementary type of data: integer, string, date, enumerated
data, etc.;
− “complex data type” is formed from several pieces of
“simple” and “complex” data types and tables. “Complex data
type”, unlike a table, does not declare an identifier in the
sense of a primary key (PK) found in the relational oriented
approach;
− a “table” is a complex type of data with a primary key
(PK), in the sense of the relational oriented approach. This
key ensures the unique identification of each occurrence of
the table.
All arrangements between the different kinds of data
structure are possible, including multi-valued information.
For example, we can model a table with columns formed by
multi-occurrence complex data type.
Finally, the term of business operation, used in semantic
modeling, is translated into a “service” at the logical
modeling level, the latter being in SOA style.
10.3. The principle of loose-coupling data
The logical architect
2
decides on the internal structure of
the domains of business objects issued from semantic
modeling. These domains are broad, composed of several
dozens of business objects, and the coupling between the
2
. The work of the logical architect has a range of action over the whole of
the information system. S/he guarantees the logical data architecture. To
respect this architecture, logical designers use derivation procedures to
establish logical data models and processing.
