7 of 8 people found the following review helpful
This review is from: Relational Database Principles (Computing Textbooks) (Paperback)
I must disagree with the other reviewers. I find the book a terrible book that is full of mistakes and misunderstood concepts.
The author confuses the 3-level architecture by writing when referring to the 3-level architecture (page 21) "On this basis, the relational model can be seen as a conceptual model". This is complete and utter rubbish, a good relational product will model all 3 levels of the 3-level architecture.
There are significant mistakes in just about every chapter. But probably the worst is chapter 5 - Normalisation. Most of the mistakes arise from a poor definition for first normal form. In the book we have the following
"The first normal form is concerned with multi-valued attributes of design (page 78)". I am not quite sure what this is supposed to mean, as it is far from clear. I would prefer something that is more precise like that given by database authorities such as Chris Date. That is, a relation is in first normal form if and only if all attributes are atomic (Single-valued). In other words, provided each attribute in any given row has a single value you have first normal form. In fact this is an essential property of any table - that is all tables are in 1NF by definition. The author attemps to normalize the data without using functional dependencies, which are barely mentioned. Instead he prefers to partition the tables by inspection, in some cases obtaining the wrong result. I have in the past used his exercises with some of my students, but have done it correctly using functional dependencies.
Not a book that I could possibly recommend. There are now many good books available for all types of student, so why waste money and time on this book. If you want a slim book that is easy to read, I would recommend:
1. The essence of databases. F.D. Rolland. Prentice Hall