I decided to learn Python recently, and I ran into "Dive Into Python" as a recommended resource (I think I Googled "Python book"), and it almost did put me off the language. It's not really the author's fault, it is the fault of those who recommend it.
I remember trying to read it way back when. It came recommended as 'the book' to read about Python.
Besides the ODBC stuff etc. that Zed comments on (never dived into the book that deeply) its biggest problem is that the writing style is extremely, annoyingly, time-wastingly redundant.
I remember it explaining the same stuff over and over, worded slightly differently, easily spanning a dozen pages for every simple thing. After the n-th "I got it the first time, move on!" feeling I stopped reading it.
I'm in the same boat. I really tried to make use of it to learn Python, but it just spends a lot of time going into detail that an interested reader should be able to infer by doing exercises.
I'm actually a big fan of that kind of writing, and I tend to write in that way myself, but it doesn't work for "intro" or "tutorial" books. Jon Skeet's "C# In Depth" is a good example of a book that spends time making sure you really understand a concept six ways from Sunday, but that's the advertised point of the book. When it comes to learning something like a new language, I think Mr. Shaw's on to something, but to be honest I'd rather see sets of exercises like this become a wikified, edited-for-quality community project. The examples in the sample PDF are geared at beginners, but there's no reason that a larger set of simple examples and drills couldn't be created by topic (here's how to do stuff with databases; here's how to ping a server; here's how to talk to a web service) for more experienced users that grok the language and are more interested in learning the libraries.
Anyways, the thing I like about this idea the most is that it puts you in front of a machine with programming tools and has you make progress by doing things, not by reading prose.
