UML in a Nutshell

Rating:
Summary: Standard "Nutshell" reference, but is this the best format?
Review: When this book came out, it was an excellent reference in an area relatively devoid of literature. Documents explaining UML (and OCL) were sparse, UML modeling tools were sparse (and proprietary), and the technology to "use" UML hadn't acheived the wide(r) acceptance and sophistication found today. However, even within this context, one might have complained about the format.

This book contains a fair bit of explanation and theory - one understands the difference between sequence and class diagrams. Moreover, one cannot really complain about the books usefulness as a reference, though the low-level "nuts-and-bolts" information falls a bit short of other O'Reilly "Nutshell" books. Yet, there is something missing in the middle. One doesn't get how to move from A (introductory this-is-this and that-is-that) to B (how do I depict relationships between classes and interfaces). In other words, one cannot read this book and begin jotting down ideas in UML - at least not easily. My largest complaint is this lack of intermediate instruction - how to move past the basics so that one can take advantage of the reference nature of the book.

However, that being said, I believe that this is a useful book to have in one's "zoo". Further, it is interesting enough to read cover to cover - I did.

Rating:
Summary: Unbelievably bad!
Review: I pity anyone who spent money on this book and thought they would get a decent reference manual to the topic. O'Reilly will lose its reputation for publishing clear, concise, readable books on technology if they keep this up. WHERE were the editors for this book? And WHO CARES whether the author "knows his stuff"? The point is for the author to communicate concepts to the reader. This abstruse work has to be the one of the greatest obfuscations of a subject I've ever seen.

After slogging through the useless preface and introduction, you get to Chapter 2: The Big Picture. You sigh with relief thinking that maybe now the author will provide some insight. This hope will evaporate quickly, sad to say. First, the author wants to enlighten you by providing an in-depth analysis of "Problems, Solutions and Problem Solving".

Let me share a couple of mangled sentences from this section:

"To deliver valued solutions (maximum quality and minimum cost within the minimum time), organizations must capture (acquire), communicate (share), and leverage (utilize) knowledge."

And later in the same section:

"In addressing this problem, an overall approach must address how we will understand or conceptualize the problem, derive a solution to the problem, and implement or realize the solution. This approach will determine how we view the problem (paradigm) for the purpose of realizing it. We will apply our knowledge of the situation and other rules of thumb (heuristics) gained from other experiences to derive the solution (artifacts). Our effort will be organized (life cycle) as a series of (possibly concurrent) steps (activities) so that it may be managed to develop the resulting information system."

Uuuhhh, and when do we get to start talking about UML? Spare us the attempts at philosophy. We know what problems and solutions are. We have a problem here -- we need to understand UML.

Oh, I can't resist adding a bit more:

"'Methods' specify how to conduct problem-solving efforts. They specify an overall problem-solving approach and its components. They specify how problems and solutions are viewed in relation to a problem-solving approach; this is known as a method's 'descriptive' aspect since it describes how knowledge is captured and communicated regarding a problem and solution. Methods also specify a problem-solving approach to be used to solve the problem and derive a solution; this is known as a methods 'prescriptive' aspect since it prescribes how knowledge is leveraged to solve a problem. Methods specify descriptively how problems and solutions are viewed, and prescriptively how the problem-solving effort may be actualized."

Do you feel edified from reading this passage? If so, run out and get this book because there's a lot more of the same throughout.



Rating:
Summary: Ouch!
Review: As a devotee of O'reilly titles (it's the first publisher I usually look to on computer subjects), this one was disappointing. Regardless of whether it covers the subject matter well, the book fails simply because of the writing. It is, at least in the beginning, weighed down by comma series that go on forever, parenthetic notes that are redundant, and points that are lost in a volume of unrelated stuff.

The author lacks focus, and the writing style is academic. This is not a practitioner's guide. I ended up purchasing a second UML book within a week.