|

Nice book! Less than 200 pages to
get you well into UML and - even if not a goal - it also teaches
about many OO techniques.

The war is over! For years that
schools taught many different modeling languages. It was a confusion!
Why did school A teach language xpto, while school B teached a different
one? Matters of taste.
But UML is a standard. Support it!
|
UML Distilled - annotated
review, part #1
UML stands for Unified Modeling Language. It is a recent standard,
and it is JUST a language... The thing is that there is some confusion
about UML, and "UML Distilled" [UMLD], a superb Martin
Fowler book, starts by trying hard to make a distinction between
a method and a JUST a language:
a method usually implies a language AND process(es), that help build
sentences with the language; but a language alone, isn't necessarily
related to a single process.
For example, think about C or PASCAL: you can write software from many
approaches, such as top-down and bottom-up techniques. On a higher level,
think about portuguese or english: you can write natural language sentences
with poetic inspiration, or with mathematical semantics...
You have a method, when you have "kind of" guidelines, to do something.
For example, there are many software engineering books that tell you
about Object Oriented [OO] methodologies, ie, ways of handling a problem
and modeling it in such a OO fashion.
Methodologies, can have multiple perspectives, and that is another regular
miss, that this book avoids. Great! Usually, Computer Science literature
takes a "conceptual", a "specification" and an "implementation" attitude,
on case studies, but authors sometimes don't stamp at what level they
are...
It is precisely on the importance of knowing your perspective, and knowing
what is right at each level, that this short book, marks a difference.
Martin Fowler writers less than 200 pages, yet he flushes you all the
key stuff, not only on UML syntax (a graphical syntax, since UML is
a graphical language, with boxes and arrows, and stuff like that), but
also on his own "informal" high level attitudes (before UML'ing),
learned from years of experience.
The book's first chapter (Introduction) writes about "use cases" - something
that is replacing the concept of "scenario" - and about "patterns", on
the context of "what is UML", and "what high level tasks should
be performed". Depending on your knowledge on "OO paths", this first
chapter has great chances of debuting you OO techniques like "CRC cards" and "interaction
diagrams"...
Overall, you have a fair chance of learning more on "UML Distilled" first
chapter's pages, than you've ever learned from those "you must act like
this, to achieve that" scholar classes, on methods. Superb! Really superb!
I don't like (I HATE!) books that cut "my personally way" [too soon],
and if there is something really unusual about Martin Fowler, it is that
he always insists that rules CAN BE BENT! and SHOULD BE BENT, as long
as you are coherent and can sustain your approach, on several layers.
After the 1st chapter, you'll have an idea about "what is UML", "what UML is
NOT", and for "what good can UML be".
|

I could have pictured the back,
but I didn't. Sorry.

Nice logo. It remembers me something
I can't quite explain...
|