UML and Data Modeling: A Reconciliation

by David C. Hay

Overview

Here you will learn how to develop an attractive, easily readable, conceptual, business-oriented entity/relationship model, using a variation on the UML Class Model notation.

This book has two audiences:

Data modelers (both analysts and database designers) who are convinced that UML has nothing to do with them; and

UML experts who don't realize that architectural data modeling really is different from object modeling (and that the differences are important).

David Hay's objective is to finally bring these two groups together in peace.

Here all modelers will receive guidance on how to produce a high quality (that is, readable) entity/relationship model to describe the data architecture of an organization. The notation involved happens to be the one for class models in the Unified Modeling Language, even though UML was originally developed to support object-oriented design. Designers have a different view of the world from those who develop business-oriented conceptual data models, which means that to use UML for architectural modeling requires some adjustments. These adjustments are described in this book.

David Hay is the author of Enterprise Model Patterns: Describing the World, a comprehensive model of a generic enterprise. The diagrams were at various levels of abstraction, and they were all rendered in the slightly modified version of UML Class Diagrams presented here. This book is a handbook to describe how to build models such as these. By way of background, an appendix provides a history of the two groups, revealing the sources of their different attitudes towards the system development process.

Product Details

ISBN-13:	9781935504191
Publisher:	Technics Publications, LLC
Publication date:	10/01/2011
Edition description:	TECHNICS PUBLICATIONS LLC
Pages:	241
Product dimensions:	7.00(w) x 10.00(h) x 0.80(d)

About the Author

Since the early 1980s, David Hay has been a pioneer in the use of process and data models to support strategic planning, requirements analysis, and system design. He has developed enterprise models for many industries, including, among others, pharmaceutical research, oil refining and production, film and television, and nuclear energy. In each case, he found the relatively simple structures hidden in formidably complex situations. Mr. Hay has published several books and numerous articles. He is a frequent speaker at professional society conferences.

Table of Contents

Chapter 2: UML and Essential Data Models 27 Impedance Mismatch 27 Architecture vs. Object-oriented Design 31 Limiting Objects to Business Objects 31 Behavior 32 Relationships and Associations 34 Entity/Relationship Predicates 36 Specifying Role Names in UML 40 A Fundamental Change to UML 42 One Solution: Stereotypes 46 Second Solution: Conversion 47 Domains, Data Types, and Enumerations 49 Namespaces 52 Object Oriented Design vs. Relational Database Design 52 Persistent Data 53 Inheritance 53 Security 55 Summary 55 Chapter 3: How to Draw an Essential Data Model in UML 57 Summary of the Approach 57 1. Show Domain-Specific Entity Cases Only 59 2. Use Symbols Selectively 60 Use Appropriate Symbols 60 Class (Entity Class) 60 Attribute 62 Association (Relationship) 63 Cardinality 65 Exclusive or (XOR) Constraint 67 Use Some UML-specific Symbols with Care 68 Entity Class Sup-types and Relationship Sub-types 69 <> 71 Derived Attributes 72 Package 74 Add One Symbol 75 Do Not Use Any Other Symbols 79 3. Define Domains 82 4. Understand “Namespaces” 84 5. Follow Display Conventions 85 Name Formats 85 Role Positions 85 “Exclusive or” Relationship Constraint 86 Cardinality Display 86 Summary 86 Chapter 4: Aesthetic guidelines and Best Practices 89 Introduction – Aesthetic Considerations 89 Place Sub-types Inside Super-types 91 Condensed Entity/Relationship Approach 91 The UML (and that of some entity/relationship notations) Approach 92 One Problem 95 Solution 95 Constraints 95 Categories 97 Eliminate Bent Lines 98 Orient “Many” End of Relationships to Top and Left 101 Presentation 102

Editorial Reviews

If you are an old-school ER modeler and now find yourself having to come up to speed on UML to get that next job (or keep the current one), this is your guidebook to success. If you are a long time object oriented programmer who has to interact with data modelers, this book is for you too. David has done the hard work of mapping out how to do a logical entity relationship model using standard (and accepted) UML diagram components. This book shows you step-by-step, with ample examples, how to get from here to there with the least pain possible for all concerned.
Kent Graziano

Certified Data Vault Master and Oracle ACE

Past-President of ODTUG & RMOUG

Brilliantly organized: three books hidden in one cohesive work. Not withstanding the tremendous value provided by cross-training data architects/modelers and object modelers/architects, making each better at what they do, Appendix B presents an absolutely awesome concise, yet detailed, history of modeling objects and data that clearly documents the differences in the approaches over the years and helps bring it all into perspective. This book is packed with useful information. Even the footnotes add clarity and offer interesting and often humorous editorial insight making it a fun read. Whatever viewpoint the reader is coming from this book has something to offer as long as the reader maintains an open mind.

Roland Berg
Senior Architect
Diligent Consulting, Inc.
San Antonio, Texas

From the Publisher