Databases Reference
In-Depth Information
These reusable semantic data models must also contain a
description of the business objects' lifecycles. This last point
will let us down. At best, data models with a good description
of business objects exist from a static point of view, but their
lifecycles are rarely available.
We will now concern ourselves with the possible principal
sources of acquiring data models: software packages,
industry specific models, and generic models.
8.5.1.
Software packages
The first source of data models is brought by software
packages (ERP, CRM, Supply Chain, etc.). With the
acquisition of a software package, a data model should be
provided. Unfortunately, software packages do not offer
sufficient transparency in terms of publication of their data
models.
Only technical documentation exists. Business
representations are weak and rarely expressed in a UML
notation.
Today, with the classic approach of software packages,
semantic modeling is too often aborted from the start. A
company finds itself trapped in its software package which
anchors itself in the Information System like a new silo. In
order for this situation to change, the software package
vendors must be interested in an approach surrounding
business repositories (MDM, BRMS and BPM) and publish
semantic models of their reference and master data, followed
by business rules. We have already discussed this new
vision, in a previous chapter (see Chapter 3). Moreover,
software package suffers from the same problems as do
“industry specific models” that we will now describe,
especially in terms of complexity which is a consequence of a
weakness in Enterprise Architecture applied to the data.
