Databases Reference
In-Depth Information
and its expressiveness. To go further, two other approaches
are possible: temporal inclusion and dynamic modeling.
Temporary inclusion
A data model can include another one. This inclusion can
be associated with a predicate, in particular a temporal one.
As long as the predicate is not valid, the initial data model is
used. When the predicate is effective, at an agreed date, the
inclusion of complementary data is triggered. The temporal
predicate is in itself master data, the initialization of which
is the responsibility of the business users.
We can, for example, couple this possibility with the
management of use contexts. With this in mind, the
initialization of the predicate is different according to the use
context of the data model. It would then be possible to
include the new part of the model at different timed stages
according to whether the use context of the initial model is a
subsidiary or the head office. Here the subsidiary and the
head office represent two examples of different contexts.
Dynamic modeling
The MDM system tool has a governance feature which
allows data to be modeled directly by authorized users, most
of the time in a test environment. On condition that this
modeling takes into account the timing aspect, that is to say
the possible period of application of the changes made to the
model, then this can be described as temporal management
applied to the data model. In this case, it would remain to be
seen how the MDM system deals with the management of
the different versions of the models.
9.3.4.3.
Transversal domain
To complement the business object domains, it is
necessary to provide for a transversal domain, accessible
from all the other domains. This domain collects re-usable
semantic models on transverse themes, like, for example,
