I tend to get too distracted to give many books the cover-to-cover treatment. Some great ones that made the cut were:
1) Metaprogramming Ruby.
2) The Art of Rails.
3) Ruby Design Patterns
4) Services-Oriented Architecture in Ruby
These stick out for me. Even if the last one in the list felt a little "rushed to production", it still covered important topics you won't find anywhere else. Plus I find when a book has typos (either in the text or in the code), that just makes me engage with it a little bit more.
> Plus I find when a book has typos (either in the text or in the code), that just makes me engage with it a little bit more.
Really? I wasted almost an hour double-checking my code due to a typo in an algorithm in _Programming Collective Intelligence_ (pg. 35, should end in "return num/den", not "return 1.0-num/den".) That's just sloppy. I added it to the errata page (http://oreilly.com/catalog/errataunconfirmed.csp?isbn=978059...).
It makes me feel that while the author's explanations are pretty good, the code samples are rather dodgy. I've switched to looking up the relevant mathematical formulas and converting them to code myself.
that book is full of errors =/ I'm not a regular python user, but I've also heard that the code contained within the book is not very "pythonic" and is really meant as a pseudocode. still no excuse for it to be incorrect.
when a book has typos, it makes me want to throw it in the fire.
I'm a bit rusty on subtleties of standard Python style these days (I mostly switched to Lua around Python 2.5 or so), but I got the same impression. Still, the conceptual explanations in the book are pretty good.
While mistaking e.g. "there" / "their" / "they're" annoys me, I acknowledge that I'm stubborn about that stuff. Messing up details in a big formula or algorithm that people are going to use as a reference is just sloppy, though. I really appreciate authors who write in a (possibly personal and ad hoc) literate programming system that automatically extracts and tests all code samples.
I don't understand your "makes me engage with it a little bit more" reaction, though. I really don't. Could you explain?
1) Metaprogramming Ruby.
2) The Art of Rails.
3) Ruby Design Patterns
4) Services-Oriented Architecture in Ruby
These stick out for me. Even if the last one in the list felt a little "rushed to production", it still covered important topics you won't find anywhere else. Plus I find when a book has typos (either in the text or in the code), that just makes me engage with it a little bit more.