I especially love the chapter on OO programming and how he breaks down each OO pattern. He starts with the most basic example of inheritance and works up to the best-case scenario, always giving the pros and cons of each pattern along the way and when each pattern could be employed. In doing so, he provides an invaluable service to those who want to understand how libraries are engineered. I remember when I first was looking at the source code for a particular library, and I was completely baffled by what I saw. For example, I would often see this:
There was no explanation to what this esoteric statement was doing. Now, after reading the book (actually, at the time it was the first edition of the book), I understand that this is known as constructor stealing or object masquerading, and I now know that it's used to inherit instance properties. I had many, many moments like this, and now when I look at source code I can intelligently follow it and understand its intent.
I also love how Zakas gives the back-history to all the subjects he covers. Knowing where stuff came from is important, even if I would never use it. For example, I never knew there were HTML methods, and if I were to have come across them in the legacy code that I support I would've thought that they were user-defined methods. Now I know better. Knowledge is power.
Lastly, the expanded chapter on Best Practices is invaluable. I've already employed them in my own work and have gotten the wheels in motion to do what I can at work (gzipping and including compression in our build process).
If you're left a little or a lot puzzled about prototypal inheritance, closures and OO programming in general, get a copy of this book and devote time to it. You'll walk away completely confident that there isn't anything out there that you can't figure out, as I did.