Three Good Things

最近偶然看到<<UML Distilled>>里，提到了我最喜欢的3个事物。

 

Since ancient times, the most talented architects and the most gifted designers have known the law 

of parsimony. Whether it is stated as a paradox ("less is more"), or a koan ("Zen mind is beginner's 

mind"), its wisdom is timeless: Reduce everything to its essence so that form harmonizes with 

function. From the pyramids to the Sydney Opera House, from von Neumann architectures to UNIX 

and Smalltalk, the best architects and designers have strived to follow this universal and eternal 

principle. 

Recognizing the value of shaving with Occam's Razor, when I architect and read I seek projects and 

books that adhere to the law of parsimony. Consequently, I applaud the book you are reading now. 

You may find my last remark surprising at first. I am frequently associated with the voluminous and 

dense specifications that define the Unified Modeling Language (UML). These specifications allow tool 

vendors to implement the UML and methodologists to apply it. For seven years, I have chaired large 

international standardization teams to specify UML 1.1 and UML 2.0, as well as several minor 

revisions in between. During this time, the UML has matured in expressiveness and precision, but it 

has also added gratuitous complexity as a result of the standardization process. Regrettably, 

standardization processes are better known for design-by-committee compromises than 

parsimonious elegance. 

What can a UML expert familiar with the arcane minutiae of the specification learn from Martin's 

distillation of UML 2.0? Quite a bit, as can you. To start with, Martin adroitly reduces a large and 

complex language into a pragmatic subset that he has proven effective in his practice. He has 

resisted the easy route of tacking on additional pages to the last edition of his book. As the language 

has grown, Martin has kept true to his goal of seeking the "fraction of UML that is most useful" and 

telling you just that. The fraction he refers to is the mythical 20 percent of UML that helps you do 80 

percent of your work. Capturing and taming this elusive beast is no mean accomplishment! 

It is even more impressive that Martin achieves this goal while writing in a wonderfully engaging 

conversational style. By sharing his opinions and anecdotes with us, he makes this book fun to read 

and reminds us that architecting and designing systems should be both creative and productive. If 

we pursue the parsimony koan to its full intent, we should find UML modeling projects to be as 

enjoyable as we found finger-painting and drawing classes in grammar school. UML should be a 

lightning rod for our creativity as well as a laser for precisely specifying system blueprints so that 

third parties can bid and build those systems. The latter is the acid test for any bona fide blueprint 

language. 

So, while this may be a small book, it is not a trivial one. You can learn as much from Martin's 

approach to modeling as you can learn from his explanations of UML 2.0. 

I have enjoyed working with Martin to improve the selection and correctness of the UML 2.0 

language features explained in this revision. We need to keep in mind that all living languages, both 

natural and synthetic, must evolve or perish. Martin's choices of new features, along with your 

preferences and those of other practitioners, are a crucial part of the UML revision process. They 

keep the language vital and help it evolve via natural selection in the marketplace. 

Much challenging work remains before model-driven development becomes mainstream, but I am 

encouraged by books like this that explain UML modeling basics clearly and apply them 

pragmatically. I hope you will learn from it as I have and will use your new insights to improve your 

own software modeling practices. 

Cris Kobryn 

Chair, U2 Partners' UML 2.0 Submission Team 

Chief Technologist, Telelogic