The Data Model Resource Book, Vol. 1: A Library of Universal Data Models for All Enterprises

Volume I contains universal models for:

- People and Organizations
- Products
- Ordering Products
- Shipments
- Work Effort
- Invoicing
- Accounting and Budgeting
- Human Resources

There is a short section on how to transform these logical data models into data warehouse models, along with some sample star schema models for Sales Analysis and Human Resources at the detailed level, with some brief examples of other star schema for the other models in the first half of the book. This title also includes about 40 pages on how to use and implement universal data models in general. While this attempts to cover at a high level the basic activities of data and process modelling along with database design, the reader should understand that the author could not possibly cover the section in any significant detail in the same book. Other texts I've recommended on this website would fill those needs.

Strengths

The first really joy of this work is that Silverston indicates at the beginning that he expects that these models will be revised, customized, and enhanced when applied to a real life problem. This obvious approach demonstrates Silverston's wealth of industry experience, over 20 years of information technology work.

As I worked my way through the models, it became obvious that these were based on data modelling efforts in the real world. The models themselves show a maturity that is rarely found in many data modelling works. There are still complexities that are not addressed, but I wouldn't be looking for those sorts of details in this type of book.

I find the inclusion of sample data values for key entities a great strength of this work.

Overall the models themselves follow, for the most part, good data modelling style and standards. The logical models are normalized, but in a practical way, not in an academic definition of normalized. What I mean by that is there are a handful of attributes that would not meet the strict criteria of normalization, but they have been denormalized in a way that is very, very common. For instance, one can find Address Line One and Address Line Two attributes in the Person and Organizations model. I don't consider this a weakness, but the author should have indicated that short cut was being used in this case.

The fact that this work includes a Party concept (and did even in its first edition) is admirable. Silverston's enhancements to the Party models since his first work show how flexible and valuable this concept can be. In the last year or so, the pendulum has swung the other way; now people can be quite irritated if a modeller does not understand the Party concept. He even includes 3 common design options for implementing Party.

Weaknesses

The book uses the Oracle/Barker notation for its diagrams and sample models. While the use of this notation is common due to the fact that Oracle modelling tools have a large installation base, its always seems odd to me that several data modelling books that hope to discuss vendor-independent and tool independent approaches to a logical data model choose to use a notation that is exclusive to one tool on the market. I don't believe there's anything wrong with the notation; I find it clear and acceptable for data modelling. I just wish that authors would choose a more commonly understood notation. A brief explanation of the notation is included in the first chapter of this work.

A common drawback of many Oracle models I see is that they follow a naming standard for relationships that leaves out the implied `to be' form of the verb in the relationship name. So instead of using "is included in" the relationship, the sample modes might use "included in". That in itself is not really a problem. The problem becomes apparent when modellers tend to use relationship names that look or act more like nouns than verbs. For instance, in the Invoice Specific Roles model, there's a relationship between SALES INVOICE and INTERNAL ORGANIZATION. This relationship is labelled "billed from" on one side "the sender of" on the other. These do not, to me, appear to be the same relationship, even if they are on the model. This sort of discrepancy can lead to modelling errors and oversights. The full verb-based approach to naming relationships would most likely use labels such as "sends" and "is sent by", clearly the same relationship. There are also several relationships that use the label "of" or "for", which to me add no more of a definition to the relationship than existence of a line between the two entities. These are not isolated occurrences; they appear to be a style of modelling used throughout the book.

Silverston's data modelling style is more informal than mine in that he does follow many common standard approaches to creating and naming modelling objects, but deviates from them in several places. He does follow, for the most part, a QUALIFIER NOUN CLASSWORD approach to naming attributes.

This brings up another weakness in this work, something that I'd hope would be changed in the next edition: these models are very tailored to American businesses. The use of the name above "Total Dollars Allowed" is not acceptable to me, even when used in a U.S. business.

These weaknesses are not significant, nor do they impact greatly on my decision to recommend this book. I'd hope that future editions would improve upon the weakness, but they are obstacles to using the models in the real world.

Please see my other review on this site concerning this included CD.


A terrific tool for the I.T. developer   June 24, 2001
Ted Kowalski (Houston, TX United States)
10 out of 11 found this review helpful

The Data Model Resource Book is as important to a data modeler or application developer as a dictionary is to a writer: you're not at your best without it. Where I've worked, this material has helped me create the architecture for an enterprise data model of our company. Aided by the Resource Book, we build a framework of major tables containing subject areas and the result is a business model to which we can map all of our application databases and data warehouses.

This book is basically rather simple to use; you find the data or subject of interest and then check to see if there are any attributes or relationships in the book that are relevant to your specific application database. This type of a check helps add quality and completeness to your logical and physical model.

But using the book just scratches the surface of its value; it's the author who's responsible for its quality and completeness. I've personally worked with Mr. Silverston who participated in consulting engagements at our firm and I`ve also seen him perform in the classroom setting. He seems to have an uncanny ability to analyze a given business situation-no matter how seemingly bizarre-and to create a model structure that will accommodate any situation.

I highly recommend the Resource Book to business analysts, application developers, programmers, and data warehouse designers.

Ted Kowalski Data Architect, Equilon Enterprises, Houston and author of "Opening Doors--A Facilitator's Handbook."


A real time-saver!   October 3, 2002
Richard Stokes (Chicago, IL United States)
9 out of 10 found this review helpful

As the Chief Architect of a major B2B manufacturing software company, I've had ample occasion to utilize this book. We were often struggling to understand the intricacies of our client's businesses so that we could model them correctly with as little frustration as possible. This book made a huge difference. I spent over 6 months modeling an industrial strength product catalog that eventually was demonstrated by Steve Ballmer at the Chicago .NET roll-out party. It did not lack for features. Yet, I could have easily reduced the development time by 3-4 months had I had this book in hand. While it didn't have *all* the answers, it would have been a significant short-cut.

Later, I used it to model office locations for a global corporation. Although our customer had originally asked us to only model US offices, I correctly anticipated that they would quickly want to expand internationally. Fortunately, this book showed me the way and the new requirement was implemented without major disruptions to the existing code.

Definitely a huge time saver and well worth the small price.


What a time saver and valuable resource!   February 26, 2004
Chandra Srinimugam (Woodinville, WA)
6 out of 6 found this review helpful

My copy of both volume 1 and volume 2 have been referenced so many times that they are tattered and worn, but such a valuable resource on all my projects. Whenever I contemplate a design, I first look at this book and, more often than not, I get ideas and best practices on ways to model the construct. I have successfully re-used designs from both volume 1 and 2 and have applied models from both of these books towards the modeling of many industries including distribution, manufacturing, telecommunications, health care, financial services and professional services.

I have found that the models provided are very relevant, re-usable, and of very high quality. They also provide all the details necessary to create physical database designs. They include entities, keys (primary and foreign), relationships (named), attributes, and the appendices include a complete listing and cross referencing of the entities and attributes, which entities are used in which diagrams (in volume 2), and domains that specify the data type and suggested lengths for each attribute. The book provides clear explanations as to why the models are modeled a certain way as well as includes data examples showing the specific instances of data that could be stored in these models.

The data models offered are at what I would consider a mid level of abstraction. Therefore, they offer a good level of flexibility without being overly abstract. The author will use very useful abstractions such as a "party", "product", "agreement" and "work effort" (with many specific subtypes of these entities as they apply to the various industries) but he does not go overboard by including overly generic concepts like "item" "activity", or "thing" entities. What I really like is that the author will often show multiple ways of modeling the same construct and show the pros and cons of the different ways to model a construct. Sometimes specific data modeling constructs are shown and sometimes more abstract entity models are shown.

Both the volume 1 common models and volume 2 industry models have not only saved me time by being able to re-use the constructs, but they have offered some perspectives that did not occur to me and have literally been a life saver to me on many a project. I could not ask for a better reference on data modeling templates and re-usable data modeling examples!


Excellent Resource!   March 9, 2004
Herman Koester (Belleville, IL USA)
6 out of 6 found this review helpful

I purchased this book along with Volume 2 a couple of years ago. The company I work for purchased the Volume 1 CD and it has been worth every penny. It has saved us countless man-hours and since our IS organization is now doing iterative development, there is no way we could ever hope to develop integrated databases without using these Universal Data Models. The book and CD is worth every cent and can benefit to any organization.

To make it clear, this is not a how-to data modeling book. It is a book of universal data models. If you need to learn how to model, purchase another book; however I would certainly recommend this book in addition to your how-to book.

Where people or organizations generally run into trouble with universal data models is when egos, not invented here syndrome, or outright pride are allowed to play a major part.

I see these universal data models analogous to the use of prehung doors and pre-assembled windows. No builder in their right mind would allow their carpenters to use their valuable time building door and window assemblies. Why would an IS manager allow his data administrators, database administrators, or developers spend their valuable time reinventing the wheel with many of these relational structures? Universal Data Models are a tremendous time saver and will allow an IS organization to develop databases faster, better, and cheaper.

I attended a conference in April 2003 and saw a presentation by a consultant, who used these Universal Data Models to create a database for Authentication and Authorization. This is another example of the powerful benefit of these models.

We have used these models on several projects with great success and will continue to use them on our upcoming projects.