Understanding the Strategy Design Pattern in Behavioral Design

1. 02 - Behavioral Design Patterns – 2 Moshe Fresko Bar-Ilan University תשס"ח 2008

2. Strategy Moshe Fresko Bar-Ilan University תשס"ו - 2005-2006 Design Patterns Course

3. Strategy • Intent • Define a family of algorithms, encapsulate each one, and make them interchangeable. Strategy lets the algorithm vary independently from clients that use it. • Motivation • Many algorithms exist for breaking a stream of text into lines. Hard-coding them in client isn’t desirable since: • Clients become more complex • Non-used algorithms will be supported • Difficult to add new algorithms • We can avoid these problems by encapsulating different line-breaking algorithms in different classes. • Each of these classes are called a strategy.

4. Strategy - Motivation

5. Strategy – Applicability • Use Strategy pattern when • Many different classes differ only in their behavior. • You need different variants of an algorithm. • An algorithm uses data that clients shouldn’t know about. • A class uses multiple conditional statements.

6. Strategy – General Structure

7. Strategy – Participants • Strategy • Declares an interface to all supported algorithms • ConcreteStrategy • Implements the algorithm using the Strategy interface • Context • Is configured with a ConcreteStrategy object • Maintains a reference to a Strategy object • May define an interface that lets the Strategy access its data.

8. Strategy • Collaborations • Strategy and Context interact to implement the chosen Algorithm. Context may pass all data required to Strategy or alternatively it can pass a reference to itself. • Clients usually create and pass a ConcreteStrategy object to the context, from a family of ConcreteStrategy objects. • Consequences • Families of related algorithsm. • An alternative to subclassing Context. • Strategies eliminate conditional statements. • A choice of implementations. Time and space trade-offs. • Clients must be aware of different strategies. • Communication overhead between Strategy and Context. • Increased number of objects. They may be shared.

9. Strategy – Implentation Issues • Implementation • Defining the Strategy and Context interfaces • Context can pass data in parameters to Strategy (better decoupling) • Context can pass itself as an argument to Strategy • Strategy can keep Context • Strategies as Template parameters • In C++ templates can be used to configure a class with a strategy. template<class Iterator> void sort(Iterator _First, Iterator _Last); template<class Iterator, class Pr> void sort(Iterator _First, Iterator _Last, Pr _Comp); • Making Strategy objects optional

10. Strategy – Sample Code import java.util.* ; class A { public final int i ; A(int i) { this.i = i ; } } class Ascending implements Comparator { public int compare(Object o1, Object o2) { A a1=(A)o1, a2=(A)o2 ; if (a1.i<a2.i) return -1 ; if (a1.i>a2.i) return +1 ; return 0 ; } } class Descending implements Comparator { public int compare(Object o1, Object o2) { A a1=(A)o1, a2=(A)o2 ; if (a1.i<a2.i) return +1 ; if (a1.i>a2.i) return -1 ; return 0 ; } }

11. Strategy – Sample Code public class S { public static void main(String[] args) { A[] asc = { new A(1), new A(5), new A(3), new A(2) } ; A[] dsc = (A[]) asc.clone() ; Arrays.sort(asc,new Ascending()) ; Arrays.sort(dsc,new Descending()) ; System.out.println("Ascending : "+arrStr(asc)) ; System.out.println("Descending: "+arrStr(dsc)) ; } static String arrStr(A[] a) { String s = "[" + a[0].i ; for (int i=1;i<a.length;++i) s += ","+a[i].i ; s+="]" ; return s ; } } // Output : Ascending : [1,2,3,5] // Descending: [5,3,2,1]

12. Strategy – Examples • class TreeSet • CTOR: TreeSet () : Default comparator. Contained objects must implement Comparable interface • CTOR: TreeSet (Comparator) • class JPanel • CTOR: JPanel () : With default layout manager • CTOR: JPanel ( LayoutManager ) • C++ STL: class map template < class Key, class Type, class Traits = less<Key>, class Allocator=allocator<pair <const Key, Type> > > class map { … }

13. Observer Moshe Fresko Bar-Ilan University תשס"ו - 2005-2006 Design Patterns Course

14. Observer • Intent • Define a one-to-many dependency between objects so that when one object changes state, all its dependents are notified and updates automatically. • Motivation • Many UI toolkits separate presentation from application data. Classes defining presentation and data can be reused independently. • Spreadsheet object and bar-chart object can depict information in the same application data. • They must be notified of any data change. • The Observer defines how to establish this relationship between “Subject” and “Observer”. • The interaction is called “Publish-Subscribe”

15. Observer - Motivation

16. Observer – Applicability • Use Observer pattern when • An abstraction has two aspects, one dependent on the other and encapsulating these aspects in different objects lets you vary and reuse them independently. • A change in an object requires changes on others and you don’t know how many objects to be changed. • An object should be able to notify others without knowing them.

17. Observer – General Structure

18. Observer – Participants • Subject • Knows its observers. • Observer • Defines an updating interface for objects that should be notified of changes in a subject. • ConcreteObject • Stores state of interest to ConcreteObserver objects. • Sends a notification to its Observers when its state changes. • ConcreteObserver • Maintains a reference to ConcreteSubject • Store state that should be consistent with the subject’s. • Implements the Observer updating interface to keep its state consistent.

19. Observer • Collaborations • ConcreteSubject notifies its observers whenever a change occurs that could make its observers' state inconsistent with its own • After being informed of a change in the concrete subject, a ConcreteObserver object may query the subject for information.

20. Observer • Consequences • Abstract coupling between Subject and Observer • Support for broadcast communication • Unexpected updates

21. Observer – Implementation • Mapping subjects to their observers • Observing more then one subject • Who triggers the update? • Subject • Client • Dangling references to deleted Subjects. • Making sure Subject state is self-consistent. • Avoiding observer-specific update protocols. • Push model • Pull model • Specifying modifications of interest explicitly.