I am disappointed that so many people have disliked
this book so intensely.

I was not the development editor for the book, but I
was the one who signed the book for publication. I
did so because I'd been trying to read UML Distilled,
and when Sinan's proposal came in, with his bullet-list
format, I found it a much quicker way to absorb the
buzzword barrage that is the price of entry to UML.

Obviously, other people didn't feel the same way (although
I'll note that along with a bunch of terrible reviews, which
echo those posted here, Amazon includes a number of
very complimentary reviews from people who had some UML
experience.

Obviously, if the book doesn't work for you guys,
it doesn't work! In this case at least, the customer
is always right. However, I do want to dispute the
idea that the book was just slapped together. It was
under development for the better part of a year, and
a good chunk of it was written before we even signed
a contract for the book.

I also want to say that Andy Oram, the developmental
editor, had some serious hesitations about the bullet-list
format, but I urged him to go ahead with it. I thought
it was an interesting experiment in how to come up with
an "In a Nutshell" compact reference for such an
abstract topic.

(As those of you who are In a Nutshell fans know, we've
extended the series from books on command-line interfaces
(UNIX in a Nutshell, Linux in a Nutshell) and programming
languages (Java in a Nutshell, Perl in a Nutshell) to all
kinds of other topics.

It hasn't always been easy to map the concept (concentrated
reference) onto material that isn't necessarily
reference oriented. As you'll see, I tried myself with
Windows 95 in a Nutshell and Windows 98 in a Nutshell
to adapt the format to a GUI-based system. I think
those books are really good books, but people from
the Windows world, who find the under-the-hood mindset
foreign, may find them a little difficult.

In a similar way, I wonder if some of the issues with this
book aren't issues with UML. It's definitely a different
way of thinking about software development.

That being said, we clearly haven't established a
reference point or a style that makes sense for a lot of people. We'll have
to go back to the drawing board.