Title: CSC 335: ObjectOriented Programming and Design
1CSC 335 Object-Oriented Programming and Design
- 27/Aug/02
- Syllabus
- Introduce Section Leaders
- Introduce JBuilder
- Object-Oriented Technology A Brief Historical
Perspective
2About this course
- Read Syllabus
- Check out the web page
- There are only two required books
- I'll ask you to read most chapters (days)
- I will enhance this material from many textbooks
that should be of interest - See recommended reading list on next slide
3A few sources of lecture enhancements during this
semester
- Computing Fundamentals with Java, Rick Mercer
- 127A book plus GUIs, events, Exceptions, IO,
designing an inheritance Hierarchy - Design Patterns Explained, A New Perspective on
Object-Oriented Design, Alan Shalloway and James
Trott - Abstract view of how design patterns can improve
software quality - Refactoring, Improving the Design of Existing
Code - Why making small changes to code results in
better code - Extreme Programming Explained, William Lake
- Concrete examples of what master programmers
- Object-Oriented Design Heuristics, Arthur Riel
- 60 guidelines to help write OO software
- Designing OO Software, Rebbeca Wirfs-Brock,
Wilkerson, Wiener - Responsibility-driven design using CRC cards
4Reading List continued
- Object-Oriented Technology OOT A Managers
Guide, Taylor - Good introduction to object-oriented technology
for IT managers. Also source for the outline and
pictures in the remaining slides.
5335 Section Leaders, Fall 2002
Section Section Leader Location Time 1
Jacob Pitts BIO E 307 800am
2 Shauna Eggers BIO E 307 900am 3H
Joshua Louie PAS 416 1100am 4
Armand Navabi BIO E 307 1200pm 5
Gergely Kota BIO E 307 100pm 6 Tom
Slattery HAURY 219 200pm 8 Cullen Linn
HARV 303 400pm 9 Charlie Chi HARV
303 500pm
6Break
- 5-minute break
- We're available for questions now.
- If you need to add or switch sections, see Rick
now
7Object-Oriented Technology Pictures from OOT A
Managers Guide, David A. Taylor
- Outline
- Consider a few ways in which data is protected
from careless modification - Mention the key features object-oriented style of
software development - Consider why object-oriented technology is
important
8Beating the Software Crisis
- Corporations continue to become more dependent on
information - Their ability to manage data decreases.
- The problem is the software, not the hardware
- The Software crisis
- How often is software delivered on time, under
budget, and does what its supposed to? - The software is not flexible enough to handle
rapid changes
9How Software is Constructed
- Wanted
- robust large-scale applications that evolve with
the corporation - It isnt easy!
- Modular Programming (the past 40 years)
- Break large-scale problems into smaller
components that are constructed independently - Programs were viewed as a collection of
procedures, each containing a sequence of
instructions
10Modular Programming
- Subroutine (1950s)
- Provided a natural division of labor
- Could be reused in other programs
- Structured Programming and Design (1960s)
- SPD was considered a good idea to program with a
limited set of control structures (no go to
statements, single returns from functions) - sequence, selection, repetition, recursion
- Program design was at the level of subroutines
- functional decomposition
11SPD Used Functional Decomposition
main
readData
saveData
mainMenu
deleteRecord
addRecord
editRecord
12A Problem with SPD
- Structured programming has a serious limitation
- Its rarely possible to anticipate the design of
a completed system before its implemented - The larger the system, the more restructuring
takes place
13And What About the Data?
- Software development had focused on the
modularization of code, - the data was either moved around between
functions via argument/parameter associations - or the data was global
- works okay for smaller programs or for big
programs when there aren't to many global
variables - Not good when variables number in the hundreds
14Dont use Global Variables
- Sharing data (global variables) is a violation of
modular programming - This makes all modules dependent on one another
- this is dangerous
- Global Data
15Information (Data) Hiding
- An improvement
- Give each subroutine its own local data
- This data can only be touched by that single
subroutine - Subroutines can be designed, implemented, and
maintained independently - Other necessary data is passed amongst the
procedures via argument/parameter associations.
16Modularized Data
- Localize data inside the modules
- This makes modules more independent of one
another. - Local Data
17Data Outside of Programs
- Small programs require little input and output
- Large programs work with the same data over and
over again - Inventory control systems
- accounting systems
- engineering design tools
- Store data in external files
18External Data
- A program that accesses data stored outside of
the program - Data stored within external files
19Sharing Data
- Many people must access the same data stored on
file - This requires a data base management system
(DBMS) - Data protected by a DBMS
20The Relational DBMS Model
- Data in a relational data base are stored in
simple tables - Each table can have primary and secondary keys
- For example, there is one table with complete
information on all customers. - Each of these records has a primary key called
custID. - MillerR 123 W. Palm, Civino, CA
- MillerT 987 E. Orange, New York, NY
- Another table stores all orders.
- Each record stores all order information with a
secondary key named customer ID. - Icees 6/8/01 MillerR 5 6.00 30.00 1.80 31.80
- Then you only need to store customer data once
- OO Systems still talk to relational data bases
21The Procedural Approach
- The procedural style of programming builds
systems one subroutine at a time. - This approach doesnt work well in large systems
- The result is defective software that is
difficult to maintain - There is a better way
22Object-Oriented Style of Design and Programming
- Three Keys to Object-Oriented Technology
- Objects
- Messages
- Classes
- Translation for structured programmers
- Variables
- Function Calls
- Data Types
23Introducing objects
- OOT began with Simula 67
- developed in Norway
- acronym for simulation language
- Why this new language?
- to build accurate models of complex working
systems - The modularization occurs at the physical object
level (not at a procedural level)
24Whats in the System
- Simula 67 was designed for system simulation (in
Norway by Kristen Nygaard and Ole-Johan Dahl) - Caller and called subprogram had equal
relationship - First notion of objects including class/instance
distinctions - Ability to put code inside an instance to be
executed - The class concept was first used here
- In procedural programming, systems are modeled as
a collection of functions and procedures that
pass data all over the place - In object-oriented programming, the system is
modeled as a collection of interacting objects
where the data is encapsulated with the methods
that need them
25Inside Objects
- An object is
- a software package that contains a collection
of related methods and data - an entity stored in computer memory
- an excellent software module
- an instance of a class
- We understand an object through
- the values the object stores (state) via
attributes (instance variables - the operations that can be applied to that object
(behavior) Booch 92.
26Modeling an Automated Vehicle
- Consider how we would mode an automated guided
vehicle (AGV) - Behaviors
- move from one location to another
- loading and unloading contents
- Must maintain information about
- its inherent characteristic pallet size, lifting
capacity, maximum speed, ... - its current state contents, location, velocity,
...
27One instance of a vehicle
- Every object has
- a name
- instance variables stored in computer memory
- methods-a.k.a. procedures, member functions, ...
28Introducing messages
- Real-world objects exhibit many effects on each
other. - These interactions are represented in software as
messages (a.ka. method or function calls). - A message has these three things
- sender the initiator of the message
- receiver the recipient of the message
- arguments data required by the receiver
29Example messages
- Smalltalk examples
- sender is the object sending the
message--sender is not shown here - vehicle104 moveTobinB7
- myAccount withdraw100.00
- Java examples
- vehicle104.moveTo( binB7 )
- myAccount.withdraw( 100.00 )
- An object-oriented program consists of objects
interacting with other objects by sending
messages to one another.
30Introducing Classes
- We often need many of the same objects within a
program--many numbers, Strings, BankAccounts,
Employees, InventoryItems, ... - We need an efficient way to redefine the same
thing many times - The class mechanism provides a template to define
the methods and instance variables of objects. - Each object (or instance) of a class has its own
state--the set of values for that instance.
31One Class can Generates Many Objects
- We can create many instances (objects) of the
same class. - Every object has
- name (a reference to it)
- state (values)
- methods
32Inheriting Class Information
- Imagine we need several classes of similar
automated Vehicles, BankAccount, or Employee
objects. - We can derive new classes from existing ones
through the inheritance mechanism. - The "top" class is called the superclass (or base
class in Java). - The new class in called a subclass (or derived
class in Java).
33Inheritance
- The superclass should pull together the common
characteristics - The subclasses inherit methods and data
- Derived classes can also add new data and methods
- Derived classes can also override some methods to
give the method new meaning - In Java you have overridden these Java methods
- toString
- equals
34Inheritance Hierarchies
- Classes can be extended to any degree.
- Each new derived class accumulates the data and
behavior (methods) of it ancestors - Much human knowledge is structured in this manner
(insects, parts, lendables in a library) - We use software models of real world objects
- The base class captures common data and behavior
35Two Small Class Hierarchies
In UML
Software needed to control these AVs
36Programming with Objects
- Simula started it all
- Smalltalk was released in early 80s
- Xerox Palo Alto Research Center PARC Place
- Alan Kay, Adele Goldberg
- It is Pure
- C started in the mid 80s
- ATT (Bjarne Stroustrup)
- Added classes to the popular C language
- Hybrid -- both procedural and object-oriented
(ugly stuff) - Ada became OO in 95
- Java started in the mid 90s just another C
extension? - C makes improvements to Java
- int, double, char are true objects (not something
different
37The OOT mindset
- Traditionally, software was developed to satisfy
a specific requirements specification. - A billing system could not be made into something
else even if were similar. - Let the billing system handle mailings or
ticklers - OOT has a different mindset
- Instead of beginning with the tasks to be
performed, OO design deals with the aspects of
the real world that need to modeled in order to
perform the tasks
38The OO approach The Good
- More natural -- models the real world, not the
computer - More flexible
- More maintainable
- later programmers understand it better
- touch as few self-contained classes as possible
- More reliable can perform unit testing
- Can build reusable modules (classes)
- Extensible
- Easier to add new functionality and offer
alternatives - Accomplished through inheritance and polymorphism
39A Wish for Reuse
- Traditional software started from scratch
- easier than converting old code--specific task
- OOT stresses reuse
- objects are the building blocks
- majority of time spent assembling proven
components ex. Graphical User Interface (GUI) - Borland's OWL, MS's MFC, or Java Swing
- But reuse is hard to obtain!
- I used to predict what we might need in the
future I was right 10 of the time A Ron
Jeffries quote during the first Birds of a
Feather on Extreme Programming (OOPSLA 1998) - Some frameworks have been successful
- MacApp, ET, Interviews, Microsoft's MFC, Java's
RMI
40The Promise of the Approach
- OOT offers
- techniques for creating flexible, natural
software modules. - systems that are much easier to adapt to new
demands - reuse shortens the development life cycle
- systems are more understandable and maintainable
- easier to remember 50 real world classes rather
than 500 functions