Home
Index Chapter 2
Chapter 4
A Twisted Look at Object Oriented Programming in C#
By Jeff Louie
09/16/2002
I must admit that my first exposure to object oriented programming (OOP) was
frustrating and difficult. As a hobbyist I have struggled through Z80 assembly
and EPROM burners, BASIC, Turbo Pascal, Java, C++ COM and now C#. The move to
event driven programming and then to object oriented programming presented major
conceptual hurdles to my function driven sequential programming mindset. The “aha”
moment when OOP made sense was most gratifying, but did not come quickly or
easily. It has been a few years since I “got” the OOP mindset and I feel
comfortable enough now to try to help fellow travelers with this journey. If OOP
comes easily to you, feel free to skip this tutorial. If you are having problems
getting your mind around objects and inheritance I hope this tutorial can help
you. This tutorial does not represent a conventional teaching method. It assumes
a passing knowledge of the C# language and familiarity with the Visual Studio
.NET IDE. This is a work in progress and may require correction or
revisions.
Comments are actively requested (email: [email protected]).
Useful Texts
I highly recommend the following books. Much of my understanding of OOP has
been gleamed from these “classic” texts and then reinforced from coding
database projects in Java, C++ and C#. At all times I willfully try to avoid
plagiarizing these authors, but my understanding of OOP is so closely tied to
these texts that I must cite them as sources of knowledge right from the
start!
Object-Oriented Analysis and Design with Applications Grady
Booch, Second Edition, Addison-Wesley, 1994, 589pp.
Design Patterns Elements of Reusable Object-Oriented Software
Gamma Helm, Johnson and Vlissides, Addison-Wesley, 1994, 395pp.
Object-Oriented Software Construction Second Edition Bertrand
Meyer, Prentice Hall, 1997, 1254pp.
Of course, some of this material is a descendent of my writing from our now
out of print book:
Visual Café for Java Explorer Database Development Edition
Brogden Louie and Tittle, Coriolis, 1998, 595pp.
Chapter 3 "Model–View/Controller"
Enough theory. It’s time to code! In this chapter, you will learn about the
most basic design pattern, the Model–View/Controller architecture (M-VC). This
lesson contains the complete source code of a Model class that encapsulates the complex math required to do
mortgage calculations. You will then create a working Mortgage calculator as a
Windows Form application. Finally, you will reuse the Model class to create a
working calculator as a Web Form application. This is all working C# code that
demonstrates the advantages of using the M-VC design pattern.
Model — View/Controller Architecture
A design pattern is a recurring solution that simplifies the design process.
The thoughtful study of design patterns allows a developer to learn from the
pioneering work of others. The M-VC design pattern is most important, I think,
because it forces a programmer to completely rethink the approach to designing
an application. The reward is reusable code. Since this is a hands on tutorial,
I plan to show you how the M-VC architecture promotes code reuse by migrating a
Windows Form solution to a browser based solution using a Web Form.
The Model — View/Controller architecture is a modification of the Model —
View — Controller architecture used in the SmallTalk language. In the SmallTalk
language, the application is divided into three parts. The Model encapsulates
the application logic or algorithms. The View draws the presentation. The Controller responds to user or system events. The key concept is
the separation of the application logic from the presentation and event handling code. The
Model class is independent of the GUI and is ignorant of any implementation of
the GUI. A good Model class should be able to function as part of a console
application and support unit testing. In this chapter, you will build an
application that separates the complex mathematics of a mortgage calculator
(Model) from the presentation and event handling code (View/Controller). The code behind
technique of Web Form programming further contributes to the separation of the
View (HTML code) and Controller (code behind event handler) code.
There is no question that the complete separation of GUI code from the
application logic is a difficult concept for many to grasp. It can be a painful
learning experience as it forces a programmer to change the basic approach to
application design. Many will fight the process, dragged into the new world of
OOP kicking and screaming. But, when you see how much simpler a Model —
View/Controller application is to maintain or migrate, you will _see_ the light.
I promise.
The Model Class
The following code was adapted from our book "Visual Cafe for Java
Explorer, Database Development Edition" Brogden, Louie, Tittel, Coriolis,
1998. The only real change from the Java version is the use of C#’s support for
"properties." The Model class implements the algorithms of a mortgage
calculator. The fact that the Java code was completely separate from the Java
GUI greatly simplified the reuse of this 1997 code in this C# Windows Form
application!
In a nutshell, given three of four mortgage parameters (principal,
interest, period in months, payment), you can calculate the unknown parameter.
Solving for interest is not a trivial mathematical solution. I readily admit
that I had help solving that equation from my good friend, Dr. W. Carlini! The
Model class has an "Init" method that takes all four parameters needed
for a mortgage calculator. One and only one parameter must have a value of zero.
The zero parameter acts as a marker for the unknown value.
This class also demonstrates the use of two common programming idioms: the
use of a public "IsValid" function to return the internal state of the
object and the use of pre and post conditions to validate the input and output
of an algorithm.
Using IsValid()
Note that all of the actual calculations occur on a call to the
"Init" method, either directly or indirectly through the "args"
constructor. If the input parameters and result appear valid, the
"target" variable is set to a valid value. If the input parameters or
result appear invalid, the "target" variable is set to -1. A public
function "IsValid" is provided to the caller that returns the internal
state of the object. The public function
"IsValid" encapsulates or hides the internal validation logic. I would argue
that the use of a public "IsValid" function is a common and useful programming idiom.
Pre and Post Conditions
The downside of separating out the algorithm from the GUI, is that both the
Model class and the View/Controller class does input checking to insure
runtime reliability and useful user feedback. The Model class implements a
common programming construct, the use of pre and post conditions. In the Model
class, the "Init" method statically validates the input parameters before passing
them on to the DoAlgorithm method (pre-conditions). The algorithm does not
check for a divide by zero error, which is handled internally. After the
calculation is complete, the "DoAlgorithm" method validates the result by calling
Double.IsNaN (IsNotANumber) (post-conditions). The decision to turn off pre and
post conditions in the release build (no check version) is beyond the scope of
this tutorial.
The compulsive coder will note the the input validation scheme is not
mathematically correct, rejecting interest rates of greater than 100%.
Apparently the twisted class has a social agenda.
Complete Code Listing Model.cs
The downside of posting all of the code is that it goes on forever. Click here
to skip the code. You can come back later.
/// <summary>
/// Class Model.cs
/// jlouie 07.07.02
/// Adapted from Model.java
/// "Visual Cafe for Java Explorer, Database Development Edition"
/// William Brogden, Jeffrey A. Louie, and Ed Tittel, Coriolis, 1998, 585pp.
/// Supplied "as is"
/// No warranty is expressed or implied
/// This code is for instructional use only
/// </summary>
public class Model
{
// internal class constants, not "versionable"
private const int INVALID= -1; // flags error
private const int PRINCIPAL= 0;
private const int INTEREST= 1;
private const int MONTHS= 2;
private const int PAYMENT= 3; private double[] arrayDb= new double[4];
private int target= INVALID;
private string message= ""; /* // uncomment to run console self test
// self test static method outputs state to console
static void ConsoleDebug(Model model)
{
if (model == null)
{
System.Console.WriteLine("Null object.");
return;
}
System.Console.WriteLine("Message: "+model.Message);
System.Console.WriteLine("Result: "+model.Result);
System.Console.WriteLine("Principal: "+model.Principal);
System.Console.WriteLine("Interest: "+model.Interest);
System.Console.WriteLine("Months: "+model.Months);
System.Console.WriteLine("Payment: "+model.Payment);
}
*/ /* // uncomment to run console self test
// self test
[STAThread]
static void Main()
{
// test internal consistency of algorithms
Model model= new Model(100000,8.5,360,0);
Model.ConsoleDebug(model); // payment = 768.9134584334
model.Init(0,8.5,360,768.9134584334);
Model.ConsoleDebug(model);
model.Init(100000,0,360,768.9134584334);
Model.ConsoleDebug(model);
model.Init(100000,8.5,0,768.9134584334);
Model.ConsoleDebug(model);
System.Console.ReadLine();
}*/
// no arg constructor
public Model(){;} // arg constructor
public Model(double principal, double interest, int months, double payment)
{
Init(principal, interest, months, payment);
} // factored code, can be called after call to constructor
// allowing reuse of instance of class
// eg. object is _not_ immutable by design
public void Init(double principal, double interest, int months, double payment)
{
// reset flags
target= INVALID;
message= ""; // store input into array of double
arrayDb[PRINCIPAL]= principal;
arrayDb[INTEREST]= interest;
arrayDb[MONTHS]= (double)months;
arrayDb[PAYMENT]= payment; // validate input // one, and only one, "value" must be zero --> target
int zeros= 0;
int tempTarget= INVALID;
for (int i=0; i<4; i++)
{
if (arrayDb[i] == 0)
{
zeros++;
tempTarget= i;
}
}
if (zeros>1)
{
message= "Too many zero parameters.";
return;
}
if (zeros == 0)
{
message= "One parameter must be zero.";
return;
} // validate interest
if (interest > 100 || interest < 0)
{
message= "Invalid interest.";
return;
} // validate months
if (months < 0)
{
message= "Invalid months.";
return;
} // validate principal
if (principal < 0)
{
message= "Invalid principal.";
return;
} // validate payment
if (payment < 0)
{
message= "Invalid payment.";
return;
} // input parameters appear valid
target= tempTarget;
DoAlgorithm(target);
} // the actual amortization algorithm
// m= P*i(1-(1+i)^-N)
// i=r/1200
// result= 0 --> marks error
private void DoAlgorithm(int target)
{
double result= 0;
double P= arrayDb[PRINCIPAL]; // principal
double i= arrayDb[INTEREST]/1200; // monthly percentage rate
double N= arrayDb[MONTHS]; // loan period in months
double m= arrayDb[PAYMENT]; // monthly payment
if (target>= 0 && target< 4) // validate target
{
try
{
switch (target)
{
case PRINCIPAL: // principal
result= 1+i;
result= 1/Math.Pow(result, N);
result= ((1-result)/i)*m;
break;
case INTEREST: // annual interest
// algorithm fails if N*m >= P !!
if ((N*m)<P)
{
throw new ArithmeticException();
}
// factor out Interest function, too long
result= CalcInterest(P,N,m);
break;
case MONTHS: // loan period
result= (1-(P*i/m));
result= Math.Log(result);
result= -result/Math.Log((1+i));
break;
case PAYMENT: // monthly payments
result= 1+i;
result= 1/Math.Pow(result,N);
result= (P*i)/(1-result);
break;
//default:
//break;
}
}
catch
{
result= 0;
}
} // validate result
if (Double.IsNaN(result))
{
result= 0;
}
if (result == 0)
{
message= "Input Error.";
}
else // valid result
{
arrayDb[target]= result;
}
}
// a complex iterative calculation for interest
// thanks to Dr. W. Carlini (and Newton)for the solution
// returns zero on error
// ASSERT (N*m)>=P
private double CalcInterest(double P, double N, double m)
{
double temp= (m/P), answer= (m/P), diff= 100, numerator= 0, denominator= 0,
accuracy= .00001;
int index, maxIterations= 1000; try
{
for (index= 0; ((diff>accuracy) && (index<maxIterations)); index++)
{
temp= answer;
numerator= (P*temp/m)+Math.Pow((1+temp),-N)-1;
denominator= (P/m)-N*Math.Pow((1+temp),(-N-1));
// if (denominator ==0 ){throw new ArithmeticException();}
answer= temp-(numerator/denominator);
diff= answer- temp;
if (diff<0)
{
diff= -diff;
}
}
answer *= 1200; // validate answer
if ((answer<0) || Double.IsNaN(answer) ||
(index == maxIterations))
{
throw new ArithmeticException();
}
}
catch
{
answer= 0;
}
return answer;
} // default target is -1 (INVALID)
public bool IsValid()
{
return ((target>=PRINCIPAL) && (target<=PAYMENT)) ? true: false;
} // Java "getter" code converted to C# get only properties
public double Result
{
get {
if (IsValid())
{
return arrayDb[target];
}
else
{
return 0.0;
}
}
}
public int Target
{
get
{
return target;
}
}
public string Message
{
get
{
return message;
}
}
public double Principal
{
get
{
if (IsValid())
{
return arrayDb[PRINCIPAL];
}
else
{
return 0.0;
}
}
}
public double Interest
{
get
{
if (IsValid())
{
return arrayDb[INTEREST];
}
else
{
return 0.0;
}
}
}
public double Months
{
get
{
if (IsValid())
{
return arrayDb[MONTHS];
}
else
{
return 0;
}
}
}
public double Payment
{
get
{
if (IsValid())
{
return arrayDb[PAYMENT];
}
else
{
return 0.0;
}
}
}
}
Top of the Model code.
Creating a Windows Form Application
I fired up the Visual Studio IDE, dragged a few controls around and wired up
two event handlers. This is what I got:
You could certainly spiff up this application by adding radio buttons that
let the user select the target of the calculation. You could then disable the
appropriate input control, setting the value of the target control to zero.
Interestingly, the Parse method happily accepts embedded commas.
In the application you simply create an instance of the Model class:
private Model model= new Model();
You then call the appropriate Model methods and properties in the
buttonCalculate and buttonReset event handlers:
private void buttonCalculate_Click(object sender, System.EventArgs e)
{
double principal= 0;
double interest= 0;
int months= 0;
double payment= 0;
bool isInputError= false;
// validate user input, must allow zero
try
{
principal= Double.Parse(textBoxPrincipal.Text);
if (principal<0)
{
throw new Exception();
}
}
catch
{
textBoxPrincipal.Text= "Invalid Input.";
isInputError= true;
}
try
{
interest= Double.Parse(textBoxInterest.Text);
if ((interest < 0) || (interest > 100))
{
throw new Exception();
}
}
catch
{
textBoxInterest.Text= "Invalid Input.";
isInputError= true;
}
try
{
months= Int32.Parse(textBoxPeriod.Text);
if (months <0)
{
throw new Exception();
}
}
catch
textBoxPeriod.Text= "Invalid Input.";
isInputError= true;
}
try
{
payment= Double.Parse(textBoxPayment.Text);
if (payment < 0)
{
throw new Exception();
}
}
catch
textBoxPayment.Text= "Invalid Input.";
isInputError= true;
}
if (isInputError)
{
return;
}
// valid user input
model.Init(principal,interest,months,payment);
if (model.IsValid())
{
textBoxPrincipal.Text= model.Principal.ToString();
textBoxInterest.Text= model.Interest.ToString();
textBoxPeriod.Text= model.Months.ToString();
textBoxPayment.Text= model.Payment.ToString();
textBoxMessage.Text= model.Message.ToString();
}
else
{
textBoxMessage.Text= model.Message.ToString();
ResetControls();
}
}
private void buttonReset_Click(object sender, System.EventArgs e)
{
textBoxMessage.Text= "";
ResetControls();
}
private void ResetControls()
{
textBoxPrincipal.Text= "";
textBoxInterest.Text= "";
textBoxPeriod.Text= "";
textBoxPayment.Text= "0";
textBoxPrincipal.Focus();
}
Creating a Web Form Application
Just to prove how easy it is to reuse the Model class, I then fired up the
IDE and built a browser based version of the mortgage calculator. Just think. If
your client suddenly wakes up one day and ask for a browser based version of
your application, you won’t be having a panic attack. By separating out the
application logic from the GUI (view and event handling code), you are prepared
for code reuse.
Here is a snapshot of the browser based calculator:
Now it’s your turn to do some coding. Just create a new Windows Form or Web
Form application and copy and paste the Model class into your project. You will
need to remove any embedded carriage returns that were added to the actual
code for easy HTML viewing. If you want to download the working projects, they are available at
http://www.geocities.com/jeff_louie/download.html
Hopefully, I have convinced you of the need to separate out your application
logic from the presentation and event handling code. I have reused the Model code from a Java application in both
a Windows Form and Web Form application. Use this most basic design pattern, the
Model — View/Controller architecture. It will grow on you!
All Rights Reserved Jeff Louie 2002
|