CSCI 370 - Lec 3: The Problem with Inheritance and the Case for Composition

---

Overview

This lecture covered several key ideas in Software Engineering, particularly focusing on:

• The pros and cons of adding more team members in software development.
• Differentiating between requirements and specifications.
• The problems associated with inheritance in object-oriented programming (OOP).
• The introduction of composition as a preferred alternative to inheritance.
• The Strategy Design Pattern as an application of composition.

---

1. Does Adding More Team Members Speed Up Software Development?

Pros:

• Workload Distribution: More people can share the workload, reducing stress and potential burnout.
• Specialization: Team members can specialize in particular tasks, improving efficiency.
• Focus: Reduced workload allows team members to concentrate on specific tasks.

Cons:

• Communication Overhead: More team members increase the need for clear instructions; miscommunication can lead to duplicate or conflicting work.
• Skill Alignment: Adding members with the same specialization leads to inefficiencies.
• Experience Level: Adding beginners may slow the team down, especially if their work interferes with advanced developers.

Conclusion: Adding team members does not always speed up development; it depends on team composition, communication, and skills.

---

2. Requirements vs Specifications

Requirements:

• Describe what the system should do.
• Aimed at stakeholders, executives, and customers.
• Example questions:
  • What can the application do?
  • What constraints does it have?
  • What are the selling points?
• Analogy: Like pitching the app to a customer or executive.
• Examples:
  • Uber Eats: allows food ordering, location tracking.
  • Games: describes available gameplay modes.

Specifications:

• Describe how the system will fulfill requirements.
• Highly technical and meant for the development team.
• Include decisions about:
  • Programming languages
  • Data structures (linked lists, hash tables, etc.)
  • Interfaces and encryption
• Example:
  • User data will be hashed for security.
  • Users will interact through drag-and-drop interfaces.

Summary:

• Requirements are promises to stakeholders.
• Specifications are instructions to the development team.

---

3. Problems with Inheritance in OOP

Inheritance:

• Mechanism to reuse code from other classes.
• Example: Class D inherits from class C to reuse method1.

Major Issues:

1. Single Inheritance in Java:

   • Java allows only single inheritance.
   • You can’t inherit from multiple classes (e.g., can’t get both A.method1 and C.method1).

2. Inheriting Unwanted Methods:

   • Subclasses inherit all methods and must override or remove unwanted ones.
   • Leads to code like empty method implementations (to disable behavior).

3. Inheritance Chain Complexity:

   • Long chains (e.g., C -> D -> E) make it hard to understand what functionality is inherited.
   • Changing a parent class can break child classes, causing fragility.

4. Tight Coupling:

   • Inheritance leads to strong coupling between parent and child.
   • A change in one class affects all descendants.

Real-World Consequences:

• Developers are hesitant to change parent classes for fear of breaking code.
• Often leads to workarounds and redundant classes.

This is known as the Fragile Base Class Problem.

---

4. Composition Over Inheritance

Composition:

• A class contains instances of other classes instead of inheriting from them.

• Offers more flexibility and decouples components.

• Example:

  class DD {
      A a = new A();
      B b = new B();
      C c = new C();
  
      void method() {
          a.method1();
          b.method2();
          c.method1();
      }
  }

Benefits:

• Choose exactly which methods to reuse.
• Avoids problems of inheritance chains.
• Easier to understand, maintain, and modify.

Principle:

Prefer Composition Over Inheritance

This is the 5th Object-Oriented Principle covered in the course.

---

5. Strategy Design Pattern

Purpose:

• Demonstrates how to use composition to vary behavior at runtime.

Problem Example:

• Different duck types: QuackingDuck, FlyingDuck, etc.
• Instead of creating a massive inheritance structure…

Solution:

• Define interfaces for behaviors:

  interface FlyBehavior {
      void fly();
  }
  
  interface QuackBehavior {
      void quack();
  }

• Implement these behaviors in concrete classes:

  class FlyHigh implements FlyBehavior {
      public void fly() { System.out.println("Flying high!"); }
  }
  
  class RegularQuack implements QuackBehavior {
      public void quack() { System.out.println("Regular quack"); }
  }

• Compose them in a Duck class:

  class Duck {
      FlyBehavior flyBehavior;
      QuackBehavior quackBehavior;
  
      Duck(FlyBehavior f, QuackBehavior q) {
          this.flyBehavior = f;
          this.quackBehavior = q;
      }
  
      void performFly() { flyBehavior.fly(); }
      void performQuack() { quackBehavior.quack(); }
  }

• In main():

  Duck d = new Duck(new FlyHigh(), new RegularQuack());
  d.performFly();    // Output: Flying high!
  d.performQuack();  // Output: Regular quack

Key Idea:

• You can switch behavior at runtime by changing the composed objects.
• Promotes flexibility and reuse without inheritance.

---

Summary

• Inheritance can lead to rigid, fragile, and hard-to-maintain code.
• Composition offers a better way to reuse code with less coupling.
• The Strategy Design Pattern illustrates how to apply composition to vary behavior dynamically.
• This lecture introduced the 5th OOP Principle: Prefer composition over inheritance.