Encapsulation

1. What Is It?

Encapsulation is the practice of bundling an object's data (fields/attributes) and the methods that operate on that data into a single unit — the class — while controlling which parts of that unit are accessible to the outside world. It is one of the four pillars of OOP alongside Abstraction, Inheritance, and Polymorphism.

Without encapsulation, any external code can freely read and modify an object's internal state. This creates invisible dependencies: a BankAccount's balance field can be set to a negative number by any caller, not just the withdraw method that enforces business rules. Encapsulation enforces that state changes flow through a controlled interface, making invariants enforceable and the class independently evolvable.

Encapsulation vs. Data Hiding: These terms are related but not synonymous. Encapsulation is the structural mechanism — bundling data and behavior together. Data hiding is a goal it enables — restricting unauthorized access to internal state. A class can be encapsulated (organized as a unit) without being tightly hidden (e.g., public fields), but strong data hiding is typically implemented through encapsulation.

2. How It Works

The core mechanics are:

1. Declare fields with a restrictive access modifier (private in Java; _ or __ prefix convention in Python).
2. Expose controlled read/write access through methods (getters/setters in Java; @property in Python).
3. Embed validation, computation, or side effects inside those accessor methods.

Mermaid Class Diagram

-+ResetFullscreen

Python

Copy
from dataclasses import dataclass


class BankAccount:
    def __init__(self, owner: str, initial_balance: float) -> None:
        self._owner = owner          # Convention: "internal use" — single underscore
        self._balance = initial_balance

    def deposit(self, amount: float) -> None:
        if amount <= 0:
            raise ValueError("Deposit amount must be positive.")
        self._balance += amount

    def withdraw(self, amount: float) -> bool:
        if amount <= 0 or amount > self._balance:
            return False
        self._balance -= amount
        return True

    @property
    def balance(self) -> float:
        """Read-only computed access — no setter exposed."""
        return self._balance

    @property
    def owner(self) -> str:
        return self._owner


# Usage
account = BankAccount("Alice", 100.0)
account.deposit(50.0)
print(account.balance)   # 150.0 — accessed via @property, not direct field
# account._balance = -999  # Possible, but violates the convention

Java

Copy
public class BankAccount {
    private final String owner;   // Immutable after construction
    private double balance;

    public BankAccount(String owner, double initialBalance) {
        this.owner = owner;
        this.balance = initialBalance;
    }

    public void deposit(double amount) {
        if (amount <= 0) {
            throw new IllegalArgumentException("Deposit amount must be positive.");
        }
        this.balance += amount;
    }

    public boolean withdraw(double amount) {
        if (amount <= 0 || amount > this.balance) {
            return false;
        }
        this.balance -= amount;
        return true;
    }

    // Getter — no setter, so balance is read-only from outside
    public double getBalance() {
        return balance;
    }

    public String getOwner() {
        return owner;
    }
}

// Usage
BankAccount account = new BankAccount("Alice", 100.0);
account.deposit(50.0);
System.out.println(account.getBalance());  // 150.0
// account.balance = -999;  // Compile error — private field

3. Variants & Comparisons

Python Access Control

Python does not enforce access at the language level. It relies on convention.

Name mangling with __attr does not create true privacy. obj._BankAccount__balance is still valid Python. Its primary use case is preventing subclass attribute collisions, not enforcing access control. For encapsulation intent, a single underscore is the Pythonic convention.

Java Access Modifiers (most to least restrictive)

Python @property vs. Java getter/setter

4. When to Use It (and When NOT To)

Use encapsulation when:

• A field has invariants that must be maintained (e.g., balance must be non-negative).
• Internal representation may change without breaking callers (e.g., storing balance in cents vs. dollars internally).
• You want to add validation, logging, or notifications on state change.
• A class's internal data structure is an implementation detail (e.g., ArrayList's backing array).

Do NOT over-encapsulate when:

• The class is a pure data container (e.g., a value object or DTO) where all fields are inherently public.
• Adding getters/setters for every field with no logic is just boilerplate — it provides zero protection and makes the code harder to read.
• Python dataclasses or Java records are the right tool: if the intent is "a named bundle of data," use those instead of hand-rolling accessors.

Anti-patterns:

• Anemic getters/setters: A class with private int x and public int getX() / public void setX(int x) with no validation is not encapsulated — it just has more verbose syntax for public fields. The field is effectively public.
• Exposing mutable internals: Returning a reference to a mutable internal collection from a getter breaks encapsulation. Return a defensive copy or an unmodifiable view instead.

Decision triggers:

• If you see "this field must always be positive" → make it private, validate in setter.
• If you see "caller shouldn't depend on how this is stored" → hide the field, expose a calculated property.
• If you see a class where every setter sets a field with no logic → reconsider whether this should be a record/dataclass with public fields.

5. Real-World Usage

1. Java's java.util.ArrayList

The backing array (transient Object[] elementData) is package-private (not public) in modern Java. Callers interact only through add(), get(), remove(), etc. This lets the JDK team resize the array, change its type, or swap the implementation entirely without breaking any code that uses ArrayList. If elementData were public, any refactor would be a breaking API change.

2. Python's @property in Django ORM models

Django model fields use property-like descriptors to intercept attribute access. When you write user.email, Django's descriptor system may lazily load, validate, or translate the value. From the caller's perspective it looks like a plain attribute — that's encapsulation enabling the ORM magic without changing the public API.

3. Java's LocalDate (immutable encapsulation)

java.time.LocalDate stores day, month, and year as private fields and provides no setters. All "modification" returns a new instance. This makes the class thread-safe by design and ensures the date always represents a valid calendar date — the constructor validates it. Encapsulation of the fields is what makes immutability enforceable.

6. Interview Cheat Sheet

Key sentences to say:

1. "Encapsulation bundles state and behavior into a class and controls access to that state through a defined interface — it's not just about making fields private, it's about ensuring state changes go through methods that can enforce invariants."
2. "Data hiding is a goal; encapsulation is the mechanism. You can have encapsulation without data hiding, but in practice they go together."
3. "Anemic getters and setters with no validation aren't encapsulation — they're public fields with extra steps. Real encapsulation means the accessor does meaningful work."
4. "In Python, privacy is enforced by convention, not the language — a single underscore signals intent. In Java, private is compiler-enforced."
5. "Encapsulation directly enables the Open/Closed Principle: by hiding the internal representation, I can change it without breaking callers."

Common follow-up questions:

Connections to other concepts:

• SRP: A well-encapsulated class tends to have a single responsibility — it owns one piece of state and all methods that operate on it.
• Open/Closed Principle: Encapsulation enables OCP — hiding internals means you can change implementation without modifying the public interface that other classes depend on.
• Dependency Inversion: Encapsulation behind an interface lets you swap implementations; DIP says you should depend on that interface, not the concrete class.
• Immutability: The strictest form of encapsulation — private fields with no setters, all state set at construction.