The book is well-written, clear, and honest to the title -- it truly is an intro. In fact, it's honest to the title a bit too much: I found it shallow. It is very much like most of the other recent C++ books (although it's one of the better-written ones), that is it has a distinct publish-or-perish taste to it, like a paper produced by another graduate student who doesn't really want to write it but has to. Not enough depth. It is, however, free from many sins of this PhD-indited flood: it's NOT pompous, it IS simple and clear, it has no pseudo-scientific pretense in it. I mean it's almost good; just not enough indepth.
Someone asked me recently, a bit confrontationally, well, you don't like anything, what's a good book then? No problem: books you tended to get a decade and more back; mostly written by practising professionals rather than CS PhD students; written by people motivated by either love of their work, or vanity, or greed -- all valid motivators, frequently resulting in good products. Unlike, I mean to say, the publish-or-perish imperative of the typical graduate student/newly minted PhD, who produce inflated and unnecessary, poorly written drivel about undeserving minutia. Abrash, Meyers, Stevens wrote good books. If you want STL, fine: Mark Nelson wrote a wonderful book on STL. It is unfortunately out of print (and behind the times a bit), but it's done right -- it really works on things, tweaks them, pokes them with a finger, looks inside, considers alternatives -- you end up really understanding the subject matter. Karlsson's book is well written, but along other books of the same kind (Josuttis, etc.) is limited to a verbal exposition of header files' contents with a teensy-weensy bit of sample code -- waaaaaay too little to be of much practical use. Whoever wants to write an STLish sorta book should check out Mark Nelson's book on STL and use it as a guide.
To summarise: The book is not bad by any means, but is superficial. Bjorn Karlsson writes very clearly, which is good and is not to be taken for granted -- and I hope Bjorn Karlsson will rewrite this book to make it more indepth, augment it with things like, you know: not only WHAT can be done, but HOW it is done (dig into the library itself: for example, how can you not want to stick your nose into the lambda library? It looks magical, I want to know how it's done... It is completely inadequate simply to mention what it can do, add a two-liner example, and be off to something else). OK, so do I recommend this book? Er... uhm... it's OK. A Quick Intro Guide, if you know what I mean. From a fifty-dollar book you'll want more.
So, I say, first go to boost and read what they've got there; I don't feel this book gets you more than the site itself -- jeez, what am I saying, of course it is less, it covers only a small part of the overall deal, but it's better written and more consistent. So, if you got fifty bucks to spare then get the book as well. I mean, it's an OK book. Were it sold for fifteen bucks, I'd give it five stars.
----------------------
PS. Bibliography is deficient: there's a couple of standard formats any style guide will describe; neither is used in here: what we have here is a kind of home-brewn summaries w/o year, publisher, etc., just the title and authors. Also, it seems that only books from Addison-Wesley made it into the bibliography (hmmm....)