I'm a software engineer at a billion-dollar company, with about 10 years of experience after college. I decided to interview at a promising early-stage startup and thus bought two interview books (this one and Programming Interview Exposed) to brush up on things and get back that interviewing magic. (I had been at my last job for over 5 years, so it has been a while.)
This book is good in some ways. The best thing about it is the set of problems in it. It's quite comprehensive: every area I could think of is covered. Even the knowledge-based chapters (e.g., on threading, or networking, or low-level programming) are surprisingly up-to-date and representative of the type of stuff that is asked. So if you want to get a big list of problems and at least a good stab at the solutions (if you're stuck), it's much better to get this book (as opposed to looking for problems on the Internet).
So why only two stars? The biggest issue is that there are tons of errors in this book. I can't tell you how many, but I know I personally found at least 5. Some are trivial -- the book may say something is O(n) space, when it's O(1) space. Some are egregious -- the smart pointer implementation, which is compressed to one page in the solution section, is utterly broken. (For example, operator=() fails to give up ownership of what it points to before the assignment, before taking ownership of the new object. That is a completely wrong solution to the problem! What if an inexperienced reader reads this section and becomes confused?) Books like this should not have mistakes, especially not a 4th edition.
Sometimes the problem isn't clearly stated, so that trying to solve it is very difficult without more explanation. Flipping to the solution reveals what the problem was, but then what was the point?
In general, when reading this book, once gets the sense that it was thrown together. It seems like it wasn't edited in the most basic ways; for example, URLs are referred to as "urls" all over the text. That doesn't really matter, of course, but it is indicative of the type of publication it is. It's like someone wrote up an elaborate text file and then basically just sent it off to be published. Very cheap and kind of unprofessional.
Finally, the book's solutions (while usually 100% correct) typically don't try to explain and expose the thinking process behind a solution. True, sometimes it's quite obvious, so there's no point. But sometimes, it's not obvious -- it would be helpful to a potential interviewee to get a glimpse into HOW to get to the solution, instead of basically just being given the solution. This isn't always the case in the book, but most of the time it is.
So, there you go. If you want a nice collection of problems that's representative and up-to-date, this is your book; but don't expect an error-free and carefully edited or explained work.
