Both books are exactly what they need to be. Nonetheless, I had the feeling that there is a market for a book that's somewhere in between: a book that explains more of the technical details than Keith's does, but not quite as much as Flanagan's does.
Therefore right now the best way to describe my book is as the sequel to Jeremy's. Once you really understand all that he explains, the time has come to immerse yourself in the more advanced features and problems of the language, as well as some nasty browser incompatibilities that can ruin your day (and your hair).
Does that mean that you should treat these three vastly different topics in the course of one chapter? On the pro side, in order to explain the script well you should explain Core, Events and DOM, because your readers need that knowledge. On the con side, if you keep doing that the book will quickly become a confused jumble of topics and subjects, where Events basics are explained in one chapter, a few more basics in another, and some advanced topics in yet another.
My solution is to strictly separate all these subjects, and to make every chapter treat all the important aspects of its topic. However, to be able to do so I had to do some serious thinking on the example scripts themselves. As I said, even the tiniest example script uses Core, Events, and DOM. How can the reader understand the entire script when the chapter he's reading only explains the Core features, but not Events and DOM?