Page 50 -

A Company and its Data 11

– the methodological and technical innovations of these
past few years, especially the object oriented approach,
process management, SOA and standards such as XML
should be made available to build rich data models, enabling
the automatic software generation of data administration
functions. This administration, as well as being business-
oriented and aligned with models, is re-named Data
Governance.

1.2.2. The rich data model

The definition of a data element requires information that
goes well beyond the description of its attributes (such as its
name, its format or its owner) that are often found in an
approach involving a data dictionary only. It is also
important to describe its relation to other data. The meaning
of the data is revealed by its associations with others. It is
not just about creating a data dictionary but a model that
expresses the relationships between the information in a
holistic fashion. Philip Howard of Bloor Research dedicates a
entire study in his report entitled “Data Discovery” and
notes that “we believe that the ability to discover and
understand the relationships that exist across your data,
wherever it resides, is of fundamental importance to a
number of IT disciplines” [HOW 09]. For instance, a product
is linked to factories that ensure its assembly and that
product is also linked to commercial entities that are
involved in its distribution. Depending on whether the
products are the responsibility of the headquarters or of
affiliate organizations, factories and commercial units can
differ. It relates to a validation which depends on the use
context in which the data is used: either by Head Office or by
affiliates. Therefore the relationships between the data
reveal validation rules that are vital to the MDM approach.
They are used by the repository to control the integrity of the
data.