Modelling Business Information

Keith Gordon

Published by

BCS, The Chartered Institute for IT


RRP £29.99
Reviewed by

David C Hay, CEO, Essential Strategies International


10 out of 10

Keith Gordon has finally created something greatly needed by the data modeling community: an introductory textbook on business data modeling (ok, “modelling”. He is British, after all…) that is both clear and disciplined. It drives home the point that analyzing business to understand it’s underlying structure is not a database function. It requires serious analytical skills to develop a business-oriented data model. Moreover, the ability to represent the results of that analysis to people not concerned with your techniques is an important discipline in its own right.

The business analytical model of an enterprise’s information represents the things of significance to that enterprise, along with rigorous (“rigourous” to Mr. Gordon) assertions about relationships among those things. The purpose of the model is to represent those assertions so that the representatives of that enterprise can assert definitively whether or not the model is a true representation of it.

The book is in two parts: The Basics and Supplementary Material.

The heart of the book (“The Basics”) is a tutorial for creating an information model that reflects the structure of an enterprise. This entails first identifying things of significance to the enterprise and the relationships among them. In addition to basic relationships, Mr. Gordon then describes more complex relationships, such as recursive relationships and mutually exclusive ones. A separate chapter is about drawing and validating diagrams. A chapter then follows that is entirely about attributes, the “information about things”.

Mr. Gordon does include a chapter about “normalization”, but that is a means to validate logic that should have already been present in the modeling exercise itself.

The “Supplementary Materials” section gives Mr. Gordon the opportunity to provide his insights on:

  • Other modeling notations: IDEF1X and Information Engineering.
  • A formal description of the rules for naming artefacts (artifacts for us Yankees) on the models: entity classes or object classes; relationships or associations; attributes and domains.
  • Three sets of criteria to evaluate data model quality, including the layout of models.
  • Corporate information and data models: This considers an enterprise-wide view of information and data, as well as corporate models.
  • Different views of data and databases - structured, unstructured and semi-structured data; master data, “Big” data, and metadata.
  • Business intelligence and related topics of data warehousing and dimensional modeling.
  • Advances in SQL, and why business analysts should not be “in the weeds”, introducing unnecessary detail too early in the process.
  • How to take a requirements information model into database design.

The book is meant to be used as a textbook, with exercises at the end of each chapter. It is not light reading, but it is both well-written and well-organized, so a serious student should appreciate the experience.

Mr. Gordon was an early student of Harry Ellis, the co-creator (with Richard Barker) of his favorite approach, which he features prominently in this book. Interestingly enough, Mr. Gordon also includes the UML notation in each of his examples. He does not take on the logic that makes most UML practitioners view the modeling world different from him, but he does adapt the notation to follow the Barker/Ellis disciplines. Each example has both Barker/Ellis and UML drawings.

Interestingly, while a glossary could reasonably be expected to be included as an appendix, Mr. Gordon has included it in the introductory pages, along with the Foreword, Preface, and Acknowledgements. What that means is that the reader starts by being immersed in the language that will be used in the book. Many of the terms should be familiar (although the definitions may not be what is expected), and others may be in the category of terms that the reader has heard but never knew what they meant. The completely mysterious terms, on the other hand, should provide good motivation for the reader to proceed with the body of the book.

In any case, anyone interested in a thoughtful, well-done text on how to do high-quality business analytical data modeling should definitely proceed with the body of this book.

Further information: BCS, The Chartered Institute for IT

August 2017