The Data Model Resource Book, Volume 3: Universal Patterns for Data Modeling

Silverston c01.tex V2 - 11/21/2008 2:57am Page 14

14 Chapter 1 ■ Introduction

accepted and specific enough to be rigorous, we have a classic ‘‘Catch-22’’

situation if we frame the discussion of patterns around these concepts. We

believe that taking a stance regarding what we consider to be a conceptual,

logical, and/or physical data model or debating the definitions of these models

would distract from what we want to offer in this book. We believe that there

is another way to categorize data models, namely by specifying the level of

generalization, and this can be more helpful in our goals.

As data modelers, we are usually asked to create data models that meet

specific business needs. For example, we are asked to create a model that

illustrates the required business data by using objects such as entities, rela-

tionships, and attributes. The enterprises that need data models want us to

create models to support particular functions, and data modelers have tried

to segment these models into categories that have meaning primarily to data

modelers (conceptual model, business model, logical model, and so on). So,

instead of using these categories we have decided to categorize data models

by how generalized the model is. In turn, the level of generalization implies

suitability of the model for a particular purpose or function. As we already

stated, very specific models are generally used to communicate