This interesting little book has now been out for twenty years. I've owned my copy for perhaps eighteen of those years. And it is interesting now to evaluate the promise suggested in this work's premise relative to the reality of the past twenty years. It is important, as well, to evaluate objectively, if I can be forgiven a slight pun, the methodology taught within the corpus of this text relative to the praxis that has emerged over the many years since its publication.
Published in 1990, Coad and Yourdon's book, here reviewed, presented a reasonable little methodology for performing object oriented analysis. What is striking today from the perspective of a professional in the field is how heavily what is presented as object oriented analysis in this work truly relies on sound principles of relational data analysis. In fact, their model of a five level schema for analysis, which would include subject, class and object, generalization/specialization, attributes, and methods/services could rightly be seen as having been completely addressed and much better dealt with in Fleming and Von Halle's classic "Handbook of Relational Database Design". The difference is that Fleming and Von Halle provided a sound methodology, well reasoned, and applicable then as now. Coad and Yourdon struggled to present something entirely different. But all they really accomplished is to validate the work earlier performed by the insightful ladies, mentioned above, who fashioned their excellent work on the sound and unimpeachable teachings of the late, great Dr. Codd.
Witness today the efficacy of the CASE tools Coad and Yourdon suggest for Object Oriented Analysis: Object Oriented Environment, OOA Tool, ObjectPlus, Adagen. Are any of these tools now market leaders? Are they even still around today? What has become of the object oriented databases the authors touted back in 1990? These included Nexpert Object, Gemstone, and Postgres. What market fate has met these technologies? And yet, DB/2 and Oracle, among other technologies supported by Fleming and Von Halle back in 1989 roll on even today. In addition, of course, ER/Win stands today as the corporate standard for data analysis. And this validates the most important truth of all: Data analysis in corporate America today proceeds from a relational paradigm. And so it will for at least the foreseeable future.
Relative to the promise and reality of the object oriented methods presented in this book, I would say now, twenty years later, that they mainly served to validate the importance of relational technology and technique. The book is useful simply because it leads the informed reader to this very important conclusion, whilst enabling him or her to converse effectively with those who continue to imbibe the objected oriented kool-aid.