Very mediocre

Regrettably, I disagree with the other reviewers here who seem to have liked this book. I didn't. Not that it's terrible: I found myself in agreement with most if not all of the rules of thumb the author talks about; the problem is more that this book is not well done.

First, this is a textbook; as such it is aimed at a student, that is someone who doesn't know the material yet and tries to learn. But for a tutorial this text overwordy and imprecise, with a generous quantity of confusing, misleading metaphors. If you happen to know what the author is trying to say, then you sorta get it -- I mean, you get it 'cause you already know it -- but a newcomer will be confused.

Sometimes the author doesn't seem to know why a particular rule is good, so he goes something like this: rule such-and-such is good because most software engineers would agree that it's good. I doubt the author conducted a statistically valid survey -- but forget that, suppose most software engineers do in fact agree: so what? All the author has now is a nice and juicy ad populum. But he, obviously, isn't even aware that something's wrong -- and that may be one of the problems with this book: the author may not be experienced enough to write books of this sort. He may be a competent practitioner, but this is not the same as teacher.

He tries to compensate via an over-abstract and smart-sounding terminology (one of the reviewers below jokingly complains that he needed to look up the word "heuristics": that is a telling comment, and it is valid. I, too, noticed that the author takes delight in using a dozen hifalutin words where one simple word would do (and result in a clearer text).

Riel also likes to belabour commonplace excessively ("data should be hidden inside objects"; "Objects should not depend on their users", blah-blah, brush teeth daily, apple a day etc.), anything to make the book thicker -- which brings me to its second defect, less important and quite a common one today: the book is blatantly padded. Out of about 400 pages only about 200 have readable text; the rest is mostly a source code printout (this code could have been supplied on an attached disk or put on the net somewhere). There's also a few pages with a concise restatement of the author's rules, which could have been printed on the cover insides (like in the Stevens network books) or supplied as a pullout; otherwise this list is hard to find and serves no purpose.

To summarize: were it better written, Heuristics could be useful for a newbie, but it's not and so it isn't. A more experienced reader, despite being annoyed, will understand it -- but only because he already knows the stuff himself and will be recognizing rather than learning. YMMV, but in my opinion this tome isn't worth reading time for any reader, new or experienced. I'm sending it back.

Other Reviews

• Lack of consistency in examples
• Should be considered a classic
• Good book but outdated
• Very mediocre
• Perfect for experienced OO developers