UML in a Nutshell

Rating:
Summary: Rare UML expertise is available here !
Review: Having studied the UML for several months on a part-time basis, I am now in possession of several books on the subject, the most recent addition being "UML In A Nutshell" by Sinan Si Alhir. The types of computer book I own could be loosely categorised under the headings of Overview, General Subject Coverage, The Practical Usage Of, Academic and Reference, and I would place "UML In A Nutshell" in the last of these. The books' three sections are: i) an excellent coverage of the meta-issues of development, such as modelling languages, problem solving and the nature of Object Orientation. ii) a concise reference covering the use of the UML. iii) a detailed description of each of the diagramming types in a quick reference format, plus supplementary information on e.g. the Object Constraint Language. Within each section, extensive use of bullet points combined with paragraphs of text and sophisticated diagrams delivers the material in a disciplined manner. Due to the nature of the UML, and the very large amount of detail in the book, it is necessary to expend considerable effort on gaining a full appreciation of the material, effort which will be repaid many times over in correct use of the language. In my view, writing as a commercial software developer, the book is excellent in its stated role - A Desktop Quick Reference. However, those intent on studying the UML must understand that the subject is MUCH more demanding than, say, making basic use of a modern programming language. This is NOT a thousand page, lightweight coverage of the latest commercial development environment, which often demand relatively little skill for basic usage. This is a concise, extremely well thought out REFERENCE work, allowing the reader access to the most subtle and powerful aspects of this sophisticated modelling language. It is ideal to supplement UML books designed for beginners, typically when they begin to ask questions of the "what does this concept actually MEAN, & how does it fit into the broader picture ?" variety. I am very glad to have added "UML In A Nutshell" to my expanding collection of texts on this fascinating subject. Well done, again, O'Reilly.

Rating:
Summary: The book should be called "A bulleted list of UML terms."
Review: I want to echo every reviewer who was commented on the writing style of this book.

I would liken trying to learn UML (the book uses 'the UML' throughout) to trying to learn a language with nothing but its dictionary. The book is almost devoid of any narrative. Instead, the author has prepared -- quite impressively -- the longest bulleted list I have ever seen.

The author's general approach is to present several pages of multi-level bulleted terms. Then he provides a diagram that attempts to illustrate all of the definitions given in earlier pages. In small doses, you might be able to penetrate the subject matter when presented this way.

But I think it would be infinitely more useful to present a situation or system, explain what needs to be documented and then show how to develop the UML diagram throughout the book.

Chapter 4, 'A Unified Modelling Language Tutorial', tries this approach. But in the totality of the book, the presentation is far too brief and the book does not stick with the sample application in the more detailed chapters (which cover the component modelling languages that UML unifies). It would have been much more useful if the imaginary project management application's diagrams were developed through the chapter.

Instead, after chapter 4, the reader is left with 150 more pages of bullet points and definitions that are NEVER tied back to that imaginary system.

Despite its dense, almost pedantic approach, it think two important things got skipped over:

* There is no coverage on exception handling within UML. It may be that UML doesn't have any way to document how a system handles when an error occurs, but I doubt it.
* The book would benefit from presenting a "palette" or "toolbox" of UML diagramming elements. For example, the significance of the # in a class diagram is mentioned in the bullet lists. A more useful approach might have been a table of valid diagramming elements and their use.

In summary, I suppose this book might be useful if you already knew UML and needed a reference for diagram types you don't normally use or to double check your understanding.

But it's not very useful if you're trying to learn UML to start documenting your own systems.

A better title for the book might be "A bulleted list of UML terms."

Rating:
Summary: Almost unreadable
Review: Every time I ... buy a technical book without checking Amazon.com for the reviews I end up kicking myself, and this time is no exception. This book is almost completely impossible to read - it is a tangle of repetitive bullet points torn from some cursed corner of academia. When the author does take the time to write a complete sentence it is riddled with odd parenthetical notes that are less than helpful. And the diagrams, at least in the initial chapters, are arcane, in my opinion rather silly, but mostly just needlessly confusing.

I found myself skipping larger and larger sections of the text at a time looking for some island of clarity from which I could learn something valuable without too much squinting and ended up skipping all the way through the book.

Of all the topics that should be treated in a clear, straightforward manner, coverage of a modeling language designed to facilitate communication pretty much tops the list - so if you are a collector of irony, this book would make a nice addition to your collection. Otherwise, I'd select another title.

Rating:
Summary: Rare UML expertise is available here !
Review: Having studied the UML for several months on a part-time basis, I am now in possession of several books on the subject, the most recent addition being "UML In A Nutshell" by Sinan Si Alhir. The types of computer book I own could be loosely categorised under the headings of Overview, General Subject Coverage, The Practical Usage Of, Academic and Reference, and I would place "UML In A Nutshell" in the last of these. The books' three sections are: i) an excellent coverage of the meta-issues of development, such as modelling languages, problem solving and the nature of Object Orientation. ii) a concise reference covering the use of the UML. iii) a detailed description of each of the diagramming types in a quick reference format, plus supplementary information on e.g. the Object Constraint Language. Within each section, extensive use of bullet points combined with paragraphs of text and sophisticated diagrams delivers the material in a disciplined manner. Due to the nature of the UML, and the very large amount of detail in the book, it is necessary to expend considerable effort on gaining a full appreciation of the material, effort which will be repaid many times over in correct use of the language. In my view, writing as a commercial software developer, the book is excellent in its stated role - A Desktop Quick Reference. However, those intent on studying the UML must understand that the subject is MUCH more demanding than, say, making basic use of a modern programming language. This is NOT a thousand page, lightweight coverage of the latest commercial development environment, which often demand relatively little skill for basic usage. This is a concise, extremely well thought out REFERENCE work, allowing the reader access to the most subtle and powerful aspects of this sophisticated modelling language. It is ideal to supplement UML books designed for beginners, typically when they begin to ask questions of the "what does this concept actually MEAN, & how does it fit into the broader picture ?" variety. I am very glad to have added "UML In A Nutshell" to my expanding collection of texts on this fascinating subject. Well done, again, O'Reilly.

Rating:
Summary: Lots of others are lots better
Review: When I write a review this bad, I feel that my objections must be as specific and concrete as possible. The good news is that the book gives me plenty to work with.

First is the writing style. Nearly all of the text is bulleted lists or hierarchies of bulleted lists. At best, it's an un-natural reading style. This isn't the best, though. Some of the lists go on for pages. Most annoying is that most bullet items start with incomplete sentences. At times, I felt desperate for a whole sentence with subject and verb.

Second is the style of illustration. I don't insist that every author be a trained graphic designer. A few basics should be obvious, though. Here's one: at least within any one picture, use different line styles to encode different information or to create contrast. Figure 2-8 uses lots of different styles to say the same thing, fig. 3-7 uses one style to say lots of different things, and fig 5-7 appears to use one line drawn in styles (weights) that change when it turns a corner. These are just examples, by the way, not a complete catalog of problem illustrations or even of kinds of problems.

The worst, though, is the technical content. The first half of the book isn't UML at all, it's an introduction to OO design (compared to two different earlier styles) and development process. I generally found the discussions so brief as to border on meaningless. If I wasn't already familiar with the topics, I doubt that I could have figured out what the machine-gun bulleted lists were trying to express. I feel sorry for the beginner trying to learn from these chapters.

Even the UML, the second half of the book, suffered from the same brevity and opaque style. The section on state charts, in particular, managed to show only the pieces but never how they fit into a whole state machine. It was as if I showed you a box full of nuts, bolts, wheels, and handles, and expected you to understand what a tricycle is. The section on activity diagrams, however, was short enough to qualify as negligent. I can't even criticize the content, because there's almost nothing there to criticize.

I can not recommend this book.

Rating:
Summary: Dense for those new to UML. A solid reference for experts
Review: The author of this In a Nutshell title clearly knows his stuff.

The Desktop Quick Reference is packed with very detailed definitions of UML concepts and the language's meta model. The author's style is very succinct and focused.

This very detailed, granular approach probably makes it a great Desktop Quick Reference for anyone who already knows UML reasonably well.

However, if you are new to UML and trying to learn it (like me), I'd say this Nutshell is pretty tough to crack. On balance, there is not a lot of narrative in the book. What is there, is very dense:

"The object-oriented paradigm is reducible to the function-driven paradigm or the data-driven paradigm. That is, while using the object-oriented paradigm, activities can be skewed to independently leverage function-driven concepts or data-driven concepts in a given problem-solving approach." (p 66)

So after several reads, the language still feels out of reach. I've gotten another book to help me teach myself: UML and the Unified Process (ASIN 0201770601).

Using the two books together, I'm making much greater strides.

And when I have learned UML, I am sure In a Nutshell will be a very useful reference.

Rating:
Summary: Not a good book
Review: This book is horrible. It tries to be a tutorial, and a reference but succeeds at neither. Instead we get a dense book consisting mostly of bullet points, complicated diagrams and a rambling format which never gets to the point or tells you concisely what you want to know. It is like someone has just cut and pasted the bullet points of their UML training course (extra complicated edition) into book format.

I don't suppose UML is *that* hard to understand assuming an author takes you through it in reasonable learning steps, but unfortunately this book doesn't.

This is a very poor O'Reilly title.

Rating:
Summary: Bad for UML and O'Reilly
Review: O'Reilly in many ways defines a technology. If there is a O'Reilly book that means it is something you probably want to know. Most of the "In A Nutshell books" have become invaluable and essential for me, but unfortunately this book misses the mark. In many ways this book does more to hurt UML than increase its use. When confronted with this confusing mess, more often than not, people will walk away thinking, UML is too confusing and not a helpful tool.

The author babbels on for two chapters without giving any concrete examples, showing the concepts, or demonstrating how it helps you. The diagrams are confusing, for example, the software life cycle diagram. I think the objective of the was author to prove he was smarter than the rest of us rather than trying to help us understand.

I want the second edition (assuming it comes out and fixes the major problems) of this book for free because it was a waste of money and is not up to O'Reilly standards. I have certainly bought enough O'Reilly books to make them quite wealthy.

Rating:
Summary: Good quick reference weak on explanation
Review: This book is more analogous to O'Reilly's *nix in a Nutshell series than to their Java in a Nutshell series. It is a very quick ready reference to the notation with some minimal explanatory material. It is only useful to practitioners. Beginners should read Fowler first, but this will then flesh out the detail Fowler misses. Alternatively get the OMG guide in PDF, it covers the same ground but in less usable electronic form.

Rating:
Summary: 273 Pages of Bullet Points
Review: UML in a Nutshell
- Not up to O'Reilly standards
- 273 Pages, nothing but bullet points
- Reads like this review
- Good thing they don't charge by the bullet
- Management people will love the bullets
- Perfect if you need bullet points for a slideshow