Information-Driven Business: How to Manage Data and Information for Maximum Advantage

Chapter 7. Describing the Enterprise

Data modeling often needs to be a labor of love, with many subtle twists and turns required to completely describe a business problem. Ted Codd defined the science and art of data modeling, starting with his paper "A Relational Model of Data for Large Shared Data Banks," as introduced in Chapter 4. The techniques continued to be developed through the 1970s and 1980s by many practitioners.

It used to be believed that it was possible to write a model for the entire organization—a so-called enterprise data model. Such a model would adhere to what's technically called third normal rules of normalization. Briefly, this means that the entities are related to each other in a unique way, avoiding the duplication of data and representing the business relationships in a completely generic way. Such an enterprise model is extremely attractive, providing flexible applications that consistently integrate. However, few businesses of any complexity have ever successfully developed such a model.

There are a number of reasons why the grand enterprise data model experiments have failed. First, to understand a data model it is necessary to understand all of the processes that use or populate the content. Such an analysis is a massive undertaking. Second, enterprise data models are an all-or-nothing affair; they do not prioritize. That is, even a minor entity can dramatically change the type of relationship. Finally, the data model is so much work that everyone assumes ...