the oo design process1.mht

VC书籍介绍C++的应用的TheOODesignProcess.zip 电子书好用的

ml#resources">Resources</A></FONT></TD></TR>
              <TR>
                <TD class=3Dtoc>&nbsp;<A=20
                  =
href=3D"http://www-106.ibm.com/developerworks/library/oo-design1/index.ht=
ml#author">About=20
                  the author</A></FONT></TD></TR>
              <TR>
                <TD bgColor=3D#0033cc height=3D5><IMG height=3D1=20
                  =
src=3D"http://www-106.ibm.com/developerworks/images/c.gif"=20
                  width=3D120></TD></TR></TBODY></TABLE></P><!-- End =
Table of Contents -->
            <BLOCKQUOTE>
              <P>Welcome to the first installment of this online class. =
My=20
              intent with this column is to provide a detailed =
experience in the=20
              object-oriented (OO) design and development process by =
actually=20
              having you do it. This column is more of a journey than an =
event,=20
              as it will take months to get through the entire process. =
We'll=20
              start out with requirements gathering, move through =
analysis to=20
              design, then do a Java implementation of that design. When =
we're=20
              finished, you'll have a complete case history of an OO =
program,=20
              literally from start to finish. I'll talk a lot about the=20
              underlying theory, but the central focus will be real =
examples of=20
              how that theory is applied.</P></BLOCKQUOTE>
            <P></P>
            <P>Before we get started in earnest next month, I've a few =
warnings,=20
            comments, and observations that will give you a feel for my =
own=20
            prejudices and opinions on the subject of design.</P><A =
name=3D1>
            <P><STRONG class=3Dsubhead>OO isn't about structure</STRONG> =

            <BR>First, at its core, object orientation is not at all =
about=20
            derivation, class hierarchies, UML, Java technology, and so =
forth.=20
            These are tools that the OO designer puts to use in order to =

            structure the analysis, the design, and the implementation, =
but they=20
            aren't central to what makes an OO program object-oriented. =
I'll=20
            certainly use all of these structural parts of object =
orientation as=20
            the process unfolds, but the first few columns in particular =
may=20
            seem mysterious to you if you equate implementation =
structure with=20
            object orientation. The key concept of object orientation is =

            <I>modeling</I>, so before we can do anything, we have to =
decide=20
            what we're going to model. Adele Goldberg (in <I>Succeeding =
with=20
            Objects</I>, see <A=20
            =
href=3D"http://www-106.ibm.com/developerworks/library/oo-design1/index.ht=
ml#resources">Resources</A>)=20
            recounts a story told to the congregation by a rabbi at the =
start of=20
            a new year:=20
            <BLOCKQUOTE>A rabbi boarded a train, one on which he often =
rode=20
              and one whose conductor, therefore, recognized him. The =
rabbi=20
              reached into his coat pocket for his ticket. Not finding =
it, he=20
              began a search of his belongings. The conductor stopped =
him:=20
              "Rabbi, I know you must have the ticket somewhere. Don't =
bother=20
              searching now. You can send it later, when you do find =
it." But=20
              the rabbi kept searching. When approached by the conductor =
yet=20
              again, the rabbi replied: "You do not understand. I know =
you trust=20
              me for the ticket, but -- where am I going?</BLOCKQUOTE>
            <P>All too many projects fail because they begin without a =
clear=20
            idea of where they're going. The OO process tries to solve =
this=20
            dilemma by defining the problem to be solved as its first =
step; it=20
            will take me a few columns to do that in sufficient detail =
so that=20
            the actual analysis and design phase can proceed. That is, =
before=20
            you can analyze, you have to have something <I>to</I> =
analyze.</P><A=20
            name=3D2>
            <P><STRONG class=3Dsubhead>Feel free to customize</STRONG> =
<BR>Next,=20
            the process I'm describing is the one that I use personally. =
I don't=20
            intend these articles to be definitive. It turns out that =
the way=20
            that I work is similar to Booch, Rumbaugh, and Jacobsen's =
RUP=20
            (Rational Unified Process, see <A=20
            =
href=3D"http://www-106.ibm.com/developerworks/library/oo-design1/index.ht=
ml#resources">Resources</A>),=20
            but that's not due to any planning on my part. Things just =
worked=20
            out that way. I also lean in the direction of Kent Beck's =
so-called=20
            Extreme Programming (XP, see <A=20
            =
href=3D"http://www-106.ibm.com/developerworks/library/oo-design1/index.ht=
ml#resources">Resources</A>)=20
            but again, this is just a matter of chance. I'd expect that =
my=20
            processes will have to be personalized by you to make them =
workable.=20
            Given that the literature is singularly devoid of a hands-on =

            discussion of any process whatsoever, it seems that a =
discussion of=20
            any process, even if it's a personal one, will be helpful. =
Please=20
            don't spend time contemplating whether I'm "right" or not; =
just bear=20
            in mind that what I do works for me, and that there's =
absolutely=20
            nothing wrong with changing what I do in a way that works =
better for=20
            you. <A name=3D3>
            <P><STRONG class=3Dsubhead>Tools, or the lack =
thereof</STRONG> <BR>A=20
            note on OO-design tools is also in order. Someone once told =
me that=20
            he couldn't do OO because he couldn't afford it: "Rose costs =

            $2500/seat and you obviously can't do design without Rose!" =
The=20
            "Rose" in question is an OO CASE (Computer-Aided Software=20
            Engineering) tool sold by Rational Software. (I may review =
OO-CASE=20
            tools in a future column.) Putting aside the fact that Rose =
is=20
            neither the most cost-effective, nor the most capable, nor =
the most=20
            usable of the tools available, I often work effectively with =
nothing=20
            more hi-tech than a pencil and paper -- no computer at =
all.</P>
            <P>As we'll see in upcoming columns, the process of OO =
design is=20
            really a form of semantic analysis. It has more in common =
with=20
            formal linguistics than traditional procedural-design =
techniques.=20
            I've found that computers often get in the way of this =
analysis. You=20
            spend more time fooling around with the tool than you spend =
getting=20
            work done. To make matters worse, many of these tools (Rose =
among=20
            them) are tied to hideous, draconian license managers that =
prevent=20
            you from running the tool on a laptop that's not connected =
via a=20
            network to the license manager. When I use a computer at =
all, I=20
            typically use a drawing program (Visio), not a CASE tool. =
(Which is=20
            not to say that I think that Visio is a particularly good =
tool for=20
            this purpose, either. I just happen to own a copy. I use my =
own=20
            templates -- see <A=20
            =
href=3D"http://www-106.ibm.com/developerworks/library/oo-design1/index.ht=
ml#resources">Resources</A>=20
            -- not the built-in UML support.)</P>
            <P>The point I'm making is that design has nothing to do =
with tools.=20
            Design is a way to crystallize your own thinking on complex =
topics,=20
            it's a way to make sure that you've covered as many bases as =

            possible before you start to code, and it's a way to =
communicate=20
            with other programmers about the code.</P>
            <P>Since communication is really central here, any tool that =
doesn't=20
            facilitate communication is virtually worthless. You are not =

            communicating if you are sitting alone in your office in =
front of a=20
            computer-based CASE tool.</P>
            <P>The choice of a CASE tool is irrelevant. I don't use any =
of them=20
            on a regular basis, though they can be helpful with larger =
projects.=20
            The downside is that they all generate miserable code, so I =
don't=20
            use the code-generation or "round-trip engineering" features =
of=20
            these tools. In any event, the notion of round-trip =
engineering --=20
            in which you create a design, generate code, hack up the =
code, then=20
            reverse-engineer the hack back into a "design" is a =
fundamentally=20
            flawed notion. Design is inherently a one-way process. You =
design,=20
            then implement from the design. You modify the design, then=20
            implement the modifications. Most of these tools create =
incorrect=20
            designs in any event, because they can't adequately reverse =
engineer=20
            the dynamic behavior of the program simply by looking at the =

            code.</P>
            <P>A decent drawing program and a word processor are more =
than=20
            sufficient for our purposes.</P><A name=3D4>
            <P><STRONG class=3Dsubhead>The design environment and =
tools</STRONG>=20
            <BR>Design is always a group activity. At a minimum, you'll =
need a=20
            programmer, a tester, a customer (or somebody who =
understands the=20
            domain of the problem thoroughly), somebody who writes =
English=20
            excellently, and a UI designer. You'll ideally have some =
sort of=20
            clerical support as well: people to take notes and create=20
            transcripts, to enter hand sketches into drawing programs, =
and so=20
            on. You can get by without this level of support, of course, =
but the=20
            design will come together more slowly since the team members =
won't=20
            be able to focus.</P>
            <P>All of these people have to be together to work =
effectively, and=20
            all of them have to be present all the time. Design is a =
full-time=20
            activity and nobody can afford to be distracted by other =
activities=20
            if you expect to finish in a timely fashion. (Virtual =
togetherness,=20
            using white-board software and the like, is a possibility, =
but=20
            you'll need video and audio feeds so you can see and hear =
everybody=20
            else, as well as the computer, if you expect to communicate=20
            effectively.)</P>
            <P>One of the best design tools you can have is a room =
designed for=20
            that purpose. My ideal design room is a pleasant room. It =
has plenty=20
            of windows and light. The walls are covered literally floor =
to=20
            ceiling with white board. (You can buy 4x8 sheets of =
Melamine at a=20
            place like Home Depot and screw them onto the wall yourself =
if you=20
            like.) This is not a conference room -- toss the conference =
table=20
            and hideously uncomfortable chairs (designed, I think, to =
make=20
            conferences as short as possible), and replace them with a=20
            comfortable couch and a few armchairs. You'll need a =
well-stocked=20
            fridge and a few small tables that can be moved where =
they're=20
            needed. The main design tool in use will be the white-board =
marker=20
            and eraser, but you'll also want an LCD projector hooked up =
to a=20
            networked machine to make it easy for the group to work =
together on=20
            written documents. Images are transferred from the white =
board to=20
            the computer with a high-res digital camera. In the best of =
all=20
            worlds, you'd combine the design room with an implementation =
room of=20
            the type described by Beck in his XP book -- add a few =
workstation=20
            clusters and some private space for people to be alone when =
they=20
            need to be.</P>
            <P>You won't find a space like this in a typical office, so =
you'll=20
            probably have to make one. Get out a screwdriver, rip up the =
cubes,=20
            and put one together. If the facilities people don't like =
it, tough.=20
            What's more important to the success of the company: =
producing=20
            high-quality software or supporting some archaic notion of =
what a=20
            "proper" office should look like?</P><A name=3D5>
            <P><STRONG class=3Dsubhead>Design and architecture</STRONG> =
<BR>One of=20