Databases Reference
In-Depth Information
3.4.2.1 Graphical Representation
Once rules have been elicited, graphical notations can promote user inter-
action. Tools that support a graphical representation make the specification
more comprehensible and appealing to the end user, thereby easing model
validation. Here, we present the extension to the E/R model presented in [4],
known as the (ER)
2
model.
As an example, consider the increase employees salary policy. This situa-
tion could be reflected using the (ER)
2
notation as shown in Figure 3.4. The
regulation is modeled as an update on the employees salary, which in turn
could lead to updating the corresponding bosss salary, if his or her salary
is below the employees. Arc labels indicate the operation to be applied to
the entity (e.g., m for modification), whereas circles and polygons stand for
events and rules, respectively. As for rules, they have a primitive event, a con-
dition over the DB state, and an action that is described as a list of DB opera-
tions or external actions. Conditions and actions are described separately
from the diagram, using textual descriptions displayed through windows
when circles or polygons are clicked on, thus avoiding cluttering of the dia-
grammatic representation [4].
3.4.3
Supporting Causal Business Policies Through Active Rules
Causal business policies concern the behavioral side of the domain. Unlike
recovering policies, now the relevant aspects to be tracked are not only state
conditions but also behavior invocation.
Lets consider the policies outlined in Section 3.3.2 and shown in
Table 3.2. The circumstances to which a policy may respond can be diverse,
namely,
r1
e2
m
e1
m
Employee
An (ER)
2
schema.
Figure 3.4
