[ifesdjeen]

I completely agree that it always depends on the goal. I'd also add a second vector: experience. If you've already worked with distributed and/or storage systems and have understanding of the field, you can jump straight to papers and code. However, for people who are just starting, it may be beneficial to read an overview book and get familiar with a field. This book attempts to be that a starting point that can help to navigate database papers and code.

Some of the folks I talked to who have already worked with databases a lot mentioned they've benefited from the book as well, since there were some concepts they were less familiar with or never had to learn (for example, transactional processing for the folks who have mostly worked with on eventually consistent databases; or B-Tree details for the folks who have mostly worked on LSM-Based storage engine)

Disclaimer: I'm author of this book.