The Data Model Resource Book: Volume 3: Universal Patterns for Data Modeling

Excellent book that lays out the options for specific models versus generalised models and the associated pros and cons. The answer is do both for different reasons and then understand how many of the business rules you want to enshrine in the data model. Builds on the first two books and shows how the authors wrote them. At the end of the day each modeller/designer must make their own calls but this provides some structure around which to make those choices. The detail wags the dog.

I like this series of books as generically they are extremely useful. I don't agree with everything in them, which actually makes them more useful as you think more rather than just accepting the pattern (or your own preference)

A belated but welcome sequel to vol.s 1-2 in the series. The book is one-of-its kind, advanced (in parts: very!), dry (in parts: very!) and will probably confuse a novice modeler. On the other hand it is practical and takes relational modeling to a new level for experienced modelers. The key point so well explained and exemplified is that there are different levels of modeling of the same subject - each with a proper purpose and use. There is never one true and correct model, a data model is after all 'a model', a representation of something, not the thing itself. The true modeling professional must know when to choose one or the other model type, and when and how to switch between them. After all, data modeling may be an art and it may be a craft, but it is definitely not an exact science (though a great many other data modeling books may leave that impression).

I bought it for completeness sake, but if you're not obsessed with not having the entire collection and the areas included in the book are not interesting to you, do not buy it.

very good book