Building a Data Model - High-Performance Web Databases: Design, Development, and Deployment - page 220

In step LDMIO, user views defined for different business functions are
combined; or perhaps different user groups are combined into one model.
For example, one user view for Ron's Real Estate Business may be defined
from the perspective of Ron, the owner, as illustrated in Exhibit 1. If Ron
also uses rental agencies to assist him in locating potential renters, a differ-
ent user view can be developed from the perspective of Clancey's Rental
Agency, dealing with multiple property owners and additional rental prop-
erties. Thus, Clancey's user view includes additional entities, subtypes,
relationships, attributes, and business rules. Presumably, in some areas,
Clancey's and Ron's user views overlap; in some areas, they may even con-
flict. In step LDMIO, the overlaps are consolidated, any inconsistencies
resolved, and new, inter-view relationships and business rules added to
form one composite logical data model.
In step LDM11, this consolidated data model is examined in light of mod-
els developed for other purposes. Again, overlaps and some inconsisten-
cies will be discovered. Most of these other models may already have been
implemented as databases; thus, the models may not be able to be changed
if errors or omissions are discovered. The objective here is to understand
and document relationships (including inconsistencies) among the designs.
This can be accomplished by comparing and defining mappings among
the data models, specifically through mappings to the business conceptual
schema. The conceptual schema evolves by combining the data models
and merging them two at a time, similar to merging user views. Then map-
pings are identified between each data model and the business conceptual
schema, including:
• differences in names
• operations performed on the conceptual schema to obtain constructs
within a particular logical data model
• interrelation of business rules
A business conceptual schema therefore allows multiple data models
and multiple database implementations that are consistent with one
another and with the business's overall operations as well as being repre-
sentative of individual user perspectives.
Finally, in step LDM12, consideration should be given to future business
changes that may affect the current data model. Those that are significant,
imminent, or probable are incorporated into, or at least documented with,
the data model. The goal is to maximize stability of the data model — to
ensure that correctness and usefulness will necessitate few changes over a
reasonable period. It is most important to incorporate any changes that
affect the business conceptual schema, because these are likely to have a
major influence on one or more individual logical data models.