From User Requirements to Conceptual Design in Data Warehouse Design - Data Warehousing Design and Advanced Engineering Applications - page 10

Figure 1. the SALE fact modeled, from the top to the bottom, using a StarER (Sapia et al., 1999), a UML
class diagram (Luján-Mora et al., 2006), and a fact schema (Hüsemann et. al, 2000)

fit their perception of the application; they only
became convinced of the necessity by following
the traces to the NFR graphs. The traceability
mechanism has also proved to be very helpful
when evaluating if a non-functional requirement
was satisfied or not since it was easier to check
the models to see what possible impacts would
arise from dropping one non-functional require-
ment or satisfying another.
To the best of our knowledge the only concep-
tual model that considers non-functional require-
ments in the area of DW is the one by Fernández-
Medina et al. (2006), that, limitedly to security
requirements, set out anAccess Control andAudit
model that allows the designer to specify access

control and audit considerations when carrying
out the conceptual modeling phase.
A different proposal comes from Peralta et al.,
(2003) that propose modelling the non-functional
requirements through guidelines that are not di-
rectly related with the conceptual model but are
instead exploited during logical design, where
most of the choices related to performance (e.g.
star vs snowflake schema, view materialization)
and security are kept.
In the rest of this section we will present the
Dimensional Fact Model (DFM) as a representa-
tive in order to give the reader a clear understand-
ing of the required expressiveness and in order
to determine the core of information that must be