1 / 44

CS 501: Software Engineering

CS 501: Software Engineering. Lecture 17 Object Oriented Design 3. Administration. Second presentation and report next week Final milestone: four weeks from handover -- design complete, implementation begun -- schedule for testing and revisions -- plan for handover to client

abram
Télécharger la présentation

CS 501: Software Engineering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CS 501: Software Engineering Lecture 17 Object Oriented Design 3

  2. Administration Second presentation and report next week Final milestone: four weeks from handover -- design complete, implementation begun -- schedule for testing and revisions -- plan for handover to client Sign up for meeting times Quiz 3 During class on Thursday Teaching Assistant Lin Guo away until April 17

  3. Design Patterns Design patterns: E. Gamma, R. Helm, R. Johnson, and J. Vlissides, Design Patterns: Elements of Reusable Object-Oriented Software. Addison-Wesley, 1994

  4. Sources The following discussion of design patterns is based on Gamma, et al., 1994, and Bruegge and Dutoit, 2004.

  5. Design Pattern Design patterns are template designs that can be used in a variety of systems. They are particularly appropriate in situations where classes are likely to be reused in a system that evolves over time. • Name. [Some of the names used by Gamma, et al. are becoming standard software terminology.] • Problem description. Describes when the pattern might be used, often in terms of modifiability and extensibility. • Solution. Expressed in terms of classes and interfaces. • Consequences. Trade-offs and alternatives.

  6. Classes can be defined in terms of other classes using inheritance. The generalization class is called the superclass and the specialization is called the subclass. If the inheritance relationship serves only to model shared attributes and operations, i.e., the generalization is not intended to be implemented, the class is called an abstract class Design for Reuse: Inheritance and Abstract Classes

  7. Implementation Inheritance Developers reuse code quickly by subclassing an existing class and refining its behavior. Is not good for reuse. Specification Inheritance The classification of concepts into type hierarchies, so that an object from a specified class can be replaced by an object from one of its subclasses. Design for Reuse: Implementation and Specification Inheritance

  8. Liskov Substitution Principle (strict inheritance) If an object of type S can be substituted in all the places where an object of type T is expected, then S is a subtype of T. Interpretation The Liskov Substitution Principle means that if all classes are subtypes of their superclasses, all inheritance relationships are specification inheritance relationships. New subclasses of T can be added without modifying the methods of T. This leads to an extensible system. Design for Reuse: Specification Inheritance

  9. Delegation A class is said to delegate to another class if it implements an operation by resending a message to another class. Delegation is an alternative to implementation inheritance that should be used when reuse is anticipated. Design for Reuse: Delegation

  10. Adapter: Wrapping Around Legacy Code Problem description: Convert the user interface of a legacy class into a different interface expected by the client, so that the client and the legacy class can work together without changes. This problem often occurs during a transitional period, when the long-term plan is to phase out the legacy system. Example: How do you use a web browser to access an existing information retrieval system that was designed for a different client?

  11. LegacyClass existingRequest() NewClass request() Adapter Design Pattern NewClient dependency OldClient

  12. ClientInterface request() Adapter request() LegacyClass existingRequest() Adapter Design Pattern: Solution Class Diagram Client abstract class delegation inheritance

  13. Adapter Design Pattern: Consequences The following consequences apply whenever the Adapter design pattern in used. • Client and LegacyClass work together without modification of either. • Adapter works with LegacyClass and all of its subclasses. • A new Adapter needs to be written if Client is replaced by a subclass.

  14. Bridge:Allowing for Alternate Implementations Name: Bridge design pattern Problem description: Decouple an interface from an implementation so that a different implementation can be substituted, possibly at runtime (e.g., testing different implementations of the same interface).

  15. Bridge: Class Diagram Client alternative implementations ConcreteImplementorA ConcreteImplementorB

  16. Bridge:Allowing for Alternate Implementations Solution: The Abstraction class defines the interface visible to the client. Implementor is an abstract class that defines the lower-level methods available to Abstraction. An Abstraction instance maintains a reference to its corresponding Implementor instance. Abstraction and Implementor can be refined independently.

  17. Bridge: Class Diagram whole/part association Client Abstraction Implementor ConcreteImplementorA ConcreteImplementorB

  18. Bridge: Class Diagram Client Abstraction Implementor RefinedAbstraction ConcreteImplementorA new abstraction ConcreteImplementorB

  19. Bridge: Consequences Consequences: Client is shielded from abstract and concrete implementations Interfaces and implementations may be tested separately

  20. Strategy:Encapsulating Algorithms Name: Strategy design pattern Problem description: Decouple a policy-deciding class from a set of mechanisms, so that different mechanisms can be changed transparently. Example: A mobile computer can be used with a wireless network, or connected to an Ethernet, with dynamic switching between networks based on location and network costs.

  21. Strategy:Encapsulating Algorithms Solution: A Client accesses services provided by a Context. The Context services are realized using one of several mechanisms, as decided by a Policy object. The abstract class Strategy describes the interface that is common to all mechanisms that Context can use. Policy class creates a ConcreteStrategy object and configures Context to use it.

  22. Strategy Example: Class Diagram for Mobile Computer Application LocationManager NetworkInterface open() close() send() NetworkConnection send() setNetworkInterface() Note the similarities to Bridge pattern Ethernet open() close() send() WirelessNet open() close() send()

  23. Strategy: Class Diagram Client Policy Context contextInterface() Strategy algorithmInterface() ConcreteStrategy1 ConcreteStrategy2

  24. Strategy: Consequences Consequences: ConcreteStrategies can be substituted transparently from Context. Policy decides which Strategy is best, given the current circumstances. New policy algorithms can be added without modifying Context or Client.

  25. Abstract Factory: Encapsulating Platforms Name: Abstract Factory design pattern Problem description: Shield the client from different platforms that provide different implementations of the same set of concepts Example: A browser must have versions that implement the same set of concepts for several windowing systems, e.g., scroll bars, buttons, highlighting, etc.

  26. Abstract Factory: Encapsulating Platforms Solution: A platform (e.g., the application for a specific windowing system) is represented as a set of AbstractProducts, each representing a concept (e.g., button). An AbstractFactory class declares the operations for creating each individual product. A specific platform is then realized by a ConcreteFactory and a set of ConcreteProducts.

  27. Abstract Factory: Class Diagram Client AbstractFactory createProductA createProductB Classes for ProductB are not shown in this diagram. There could be several ConcreteFactory classes, each a subclass of AbstractFactory, ConcreteFactory1 createProductA createProductB AbstractProductA ProductA

  28. Abstract Factory: Consequences Consequences: Client is shielded from concrete products classes Substituting families at runtime is possible Adding new products is difficult since new realizations must be created for each factory

  29. Command:Encapsulating Control Flow Name: Command design pattern Problem description: Encapsulates requests so that they can be executed, undone, or queued independently of the request. Solution: A Command abstract class declares the interface supported by all ConcreteCommands. ConcreteCommands encapsulate a service to be applied to a Receiver. The Client creates ConcreteCommands and binds them to specific Receivers. The Invoker actually executes a command.

  30. Command: Class Diagram invokes Command execute() Invoker ConcreteCommand1 execute() <<binds>> Receiver ConcreteCommand2 execute()

  31. Command: Class Diagram for Match invokes Move play() replay() Match Game1Move play() replay() <<binds>> GameBoard Game1Move play() replay()

  32. Command: Consequences Consequences: The object of the command (Receiver) and the algorithm of the command (ConcreteCommand) are decoupled. Invoker is shielded from specific commands. ConcreteCommands are objects. They can be created and stored. New ConcreteCommands can be added without changing existing code.

  33. Composite:Representing Recursive Hierarchies Name: Composite design pattern Problem description: Represents a hierarchy of variable width and depth, so that the leaves and composites can be treated uniformly through a common interface. Solution: The Component interface specifies the services that are shared between Leaf and Composite. A Composite has an aggregation association with Components and implements each service by iterating over each contained Component. The Leaf services do the actual work.

  34. Composite: Class Diagram Client Component * leaves Leaf Composite

  35. Composite: Consequences Consequences: Client uses the same code for dealing with Leaves or Composites. Leaf-specific behavior can be changed without changing the hierarchy. New classes of Leaves can be added without changing the hierarchy.

  36. Facade:Encapsulating Subsystems Name: Facade design pattern Problem description: Reduce coupling between a set of related classes and the rest of the system. Solution: A single Facade class implements a high-level interface for a subsystem by invoking the methods of the lower-level classes. Example. A Compiler is composed of several classes: LexicalAnalyzer, Parser, CodeGenerator, etc. A caller, invokes only the Compiler (Facade) class, which invokes the contained classes.

  37. Facade: Class Diagram Facade Facade service() Class1 service1() Class2 service2() Class3 service3()

  38. Facade: Consequences Consequences: Shields a client from the low-level classes of a subsystem. Simplifies the use of a subsystem by providing higher-level methods. Enables lower-level classes to be restructured without changes to clients. Note. The repeated use of Facade patterns yields a layered system.

  39. Observer:Encapsulating Control Flow Name: Observer design pattern Problem description: Maintains consistency across state of one Subject and many Observers. Solution: A Subject has a primary function to maintain some state (e.g., a data structure). One or more Observers use this state, which introduces redundancy between the states of Subject and Observer. Observer invokes the subscribe() method to synchronize the state. Whenever the state changes, Subject invokes its notify() method to iteratively invoke each Observer.update() method.

  40. Observer: Class Diagram subscribers Subject subscribe() unsubscribe() notify() Observer update() * 1 ConcreteSubject state getstate() setstate() ConcreteObserver observeState update()

  41. Observer: Consequences Consequences: Decouples Subject, which maintains state, from Observers, who make use of the state. Can result in many spurious broadcasts when the state of Subject changes.

  42. Proxy:Encapsulating Expensive Objects Name: Proxy design pattern Problem description: Improve performance or security of a system by delaying expensive computations, using memory only when needed, or checking access before loading an object into memory. Solution: The ProxyObject class acts on behalf of a RealObject class. Both implement the same interface. ProxyObject stores a subset of the attributes of RealObject. ProxyObject handles certain requests, whereas others are delegated to RealObject. After delegation, the RealObject is created and loaded into memory.

  43. Proxy: Class Diagram Client Object filename op1() op2() ProxyObject filename op1() op2() 0..1 RealObject data:byte[] op1() op2() 1

  44. Proxy: Consequences Consequences: Adds a level of indirection between Client and RealObject. The Client is shielded from any optimization for creating RealObjects.

More Related