This is a course book, with objectives, conceptual summaries and exercises at the end of each chapter. Publishing it here is a way of saying something that rarely gets said: part of what anyone reads is study material. That is neither good nor bad, it is a tool for passing a subject, and it is worth telling apart from a book that changes how you work.
What survived
The three-level schema architecture and the logical and physical independence of data. It sounds like 1970s theory, it is 1970s theory, and it explains a good share of what breaks today: a report that reads tables directly is tied to the physical shape of the data, so it falls over every time someone underneath moves something. It is the argument I use for putting a view or an intermediate model in place even when it looks like an unnecessary detour.
The separation of design stages, with loading, operation and maintenance treated as stages in their own right. Most projects that go wrong skipped the conceptual stage, the one where you are still arguing about what each thing means before it becomes a table.
What I left out
Ninety pages of worked queries, which are classroom drills that any tool writes today. And the chapter on object-oriented databases with a section on NoSQL and Cassandra, which was already late in 2017 and now reads as a period piece. The book is honest about what it is and claims nothing more: anyone looking for a text to decide data architecture in a company has to look elsewhere.