System Design (HLD)Encapsulation — Protecting Data & Behavior

Data Hiding Principles

LevelBeginner

Duration60 mins

TopicEncapsulation — Protecting Data & Behavior

1 / 4

Private Fields, Public Interface

The Invisible Barrier That Makes Software Work

Every well-designed software component shares a fundamental characteristic: it hides more than it shows. Like an iceberg where the visible tip represents only a fraction of its total mass, a well-designed class exposes a small, carefully crafted interface while concealing the vast complexity of its internal implementation.

This isn't accidental. It's the result of a deliberate design principle that has proven itself across decades of software engineering: private fields, public interface. This principle is the operational foundation of encapsulation—the first pillar of object-oriented design.

Understanding this principle deeply will transform how you think about class design. You'll learn to see the boundary between public and private not as a mere syntax requirement, but as a fundamental architectural decision that determines whether your software can evolve gracefully or collapses under its own complexity.

What You Will Master

By the end of this page, you will understand why private fields are the default, how to design clean public interfaces, the relationship between visibility and coupling, and how this principle enables software that can change without breaking. These concepts form the bedrock of professional software design.

The Fundamental Principle

The principle can be stated simply:

Make all fields private. Expose only what clients genuinely need through carefully designed public methods.

This sounds straightforward, yet it's consistently violated by developers at all experience levels. To truly understand why this principle matters, we need to examine what happens when it's followed—and what happens when it's ignored.

Private Fields Enable

•Implementation Freedom — Change internal representation without affecting clients
•Invariant Protection — Guarantee that internal state remains valid
•Reduced Coupling — Clients depend only on behavior, not structure
•Controlled Evolution — Add features without breaking existing code
•Simplified Reasoning — Understand state changes through explicit methods

Public Fields Cause

•Implementation Lock-in — Any change breaks client code
•Invariant Violations — No way to enforce valid state
•Tight Coupling — Clients depend on internal structure
•Fragile Evolution — Every change risks cascading failures
•Scattered Logic — Validation and rules spread everywhere

The key insight: The distinction between private and public isn't about secrecy—it's about defining the contract between your class and the world. Private members are your implementation details; public members are your promises.

When you make something public, you're making a commitment. You're saying: "This will continue to exist and work the same way. You can depend on it." When you make something private, you're reserving the right to change it. You're saying: "This is how I've chosen to solve the problem today, but I may solve it differently tomorrow."

Anatomy of a Well-Designed Class

Let's examine what a properly encapsulated class looks like by contrasting two approaches to the same problem: representing a bank account with balance constraints.

The Problem: Design a class representing a bank account that:

Has a balance that cannot go negative
Tracks the account holder's name
Records when the account was opened

BankAccount — Poor Design (Public Fields)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
// ❌ POOR DESIGN: Public fields expose implementation details
public class BankAccount {
    public double balance;           // Anyone can modify!
    public String holderName;        // Anyone can change!
    public Date openedDate;          // Anyone can alter history!
    
    public BankAccount(String holderName, double initialBalance) {
        this.holderName = holderName;
        this.balance = initialBalance;
        this.openedDate = new Date();
    }
}
 
// Client code (anywhere in the system):
BankAccount account = new BankAccount("Alice", 1000.0);
account.balance = -500.0;           // Negative balance! No protection.
account.holderName = "";            // Empty name! No validation.
account.openedDate = new Date(0);   // Altered history! No integrity.
 
// The class cannot protect its own invariants.
// Business rules must be enforced everywhere this class is used.

Now contrast this with a properly encapsulated design:

BankAccount — Good Design (Private Fields)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
// ✅ GOOD DESIGN: Private fields with controlled access
public class BankAccount {
    private double balance;
    private String holderName;
    private final Date openedDate;  // Immutable after construction
    
    public BankAccount(String holderName, double initialBalance) {
        validateHolderName(holderName);
        validateBalance(initialBalance);
        
        this.holderName = holderName;
        this.balance = initialBalance;
        this.openedDate = new Date();
    }
    
    // ===== PUBLIC INTERFACE: What clients can do =====
    
    public double getBalance() {
        return balance;
    }
    
    public String getHolderName() {
        return holderName;
    }
    
    public Date getOpenedDate() {
        return new Date(openedDate.getTime());  // Defensive copy!
    }
    
    public void deposit(double amount) {
        if (amount <= 0) {
            throw new IllegalArgumentException("Deposit must be positive");
        }
        balance += amount;
    }
    
    public void withdraw(double amount) {
        if (amount <= 0) {
            throw new IllegalArgumentException("Withdrawal must be positive");
        }
        if (amount > balance) {
            throw new IllegalStateException("Insufficient funds");
        }
        balance -= amount;
    }
    
    // ===== PRIVATE IMPLEMENTATION: How invariants are enforced =====
    
    private void validateHolderName(String name) {
        if (name == null || name.trim().isEmpty()) {
            throw new IllegalArgumentException("Holder name required");
        }
    }
    
    private void validateBalance(double amount) {
        if (amount < 0) {
            throw new IllegalArgumentException("Initial balance cannot be negative");
        }
    }
}
 
// Client code is now FORCED to go through the interface:
BankAccount account = new BankAccount("Alice", 1000.0);
account.deposit(500.0);    // ✓ Valid operation
account.withdraw(200.0);   // ✓ Valid operation
account.withdraw(5000.0);  // ✗ Throws exception - invariant protected!

The Critical Difference

In the well-designed version, it's impossible to create an invalid BankAccount or put it into an invalid state through normal operations. The class is self-protecting. Clients don't need to remember to validate; the class enforces its own rules. This is the essence of encapsulation.

The Public Interface as Contract

When you expose something publicly—whether a method, property, or constant—you're entering into a contract with every piece of code that will ever use your class. This contract has profound implications:

The Contract Promises:

Existence — The public member will continue to exist in future versions
Signature Stability — The parameters and return type won't change incompatibly
Behavioral Consistency — The method will continue to do what it currently does
Semantic Integrity — The meaning of the member won't fundamentally change

Breaking any of these promises breaks client code. In a large system, a single broken promise can cascade into hundreds of compilation errors or—far worse—silent behavioral changes that take weeks to diagnose.

Contract Implications of Public vs Private
Aspect	Public Member	Private Member
Change Freedom	Nearly frozen—changes break clients	Complete freedom to refactor
Documentation Burden	Must be documented for external use	Internal comments sufficient
Testing Requirements	Must test for backward compatibility	Test for correctness only
Version Coupling	Clients lock to your version	Clients unaffected by changes
Design Pressure	Must be near-perfect upfront	Can iterate and improve freely

The asymmetry is stark: Making something private gives you complete freedom; making something public chains you to that decision forever.

This is why the advice is always "start private, go public only when necessary." It's vastly easier to make a private member public later (if clients genuinely need it) than to make a public member private (which breaks all existing clients).

The Minimum Interface Principle:

Expose the smallest public interface that enables clients to accomplish their goals. Every additional public member is additional maintenance burden and additional constraint on your future evolution.

Ask yourself for each potential public member: "Does the client need this, or is it just convenient?" Convenience is not sufficient justification for public exposure.

The Ratchet Effect

Once something is public and clients depend on it, it's almost impossible to remove. Every public API decision is effectively permanent. This is why library and framework designers obsess over their public interfaces—they know they'll be living with those decisions for decades.

Designing the Public Interface

A well-designed public interface has specific characteristics that distinguish it from a collection of exposed fields. Understanding these characteristics helps you craft interfaces that serve clients well while preserving your implementation freedom.

Characteristics of Excellent Public Interfaces

•Behavior-Centric — Express what clients can do, not what data you hold. account.deposit(amount) rather than account.balance += amount.
•Self-Documenting — Method names clearly communicate purpose and intent. A reader shouldn't need to inspect implementation to understand usage.
•Minimal — Include only what clients genuinely need. Resist the urge to expose 'just in case' features.
•Consistent — Follow naming conventions and patterns uniformly. If one method is getUserById(), don't have another called fetchOrderForId().
•Complete — Enable all legitimate use cases without requiring workarounds. If clients can't accomplish their goals, they'll resort to hacks.
•Orthogonal — Each method does one thing. Avoid overlapping functionality that can be composed from other public methods.

The Abstraction Level Rule:

Your public interface should operate at a higher abstraction level than your private implementation. This is what allows the interface to remain stable while the implementation evolves.

Consider the difference:

Abstraction Levels in Interface Design
Low Abstraction (Poor)	High Abstraction (Good)
`setBalanceField(double value)`	`deposit(double amount)`
`getInternalHashMap()`	`findUserById(String id)`
`sqlQuery(String query)`	`findActiveOrders()`
`setDatabaseConnection(Connection c)`	`connect(String url)`
`getArrayList()`	`getItems()` returning List interface

The low-abstraction versions expose implementation details: field names, internal data structures, database specifics. Changing any of these requires changing the interface.

The high-abstraction versions describe intent: what the client wants to accomplish. The implementation can change freely—use a different data structure, switch databases, refactor the internal representation—without affecting the interface.

The Shopping Cart Test

When designing a public method, imagine you're an engineer at a different company who's never seen your code. Read only the method signature and documentation. Can you understand exactly what the method does and how to use it? If you need to see the implementation to understand the interface, the abstraction level is too low.

Common Anti-Patterns in Field Visibility

Even experienced developers fall into visibility anti-patterns. Recognizing these patterns in your own code—and understanding why they're problematic—is essential to applying the private fields principle correctly.

Anti-Pattern 1: The Anemic Domain Model

•What it is: Classes with all private fields but trivial getters and setters for everything, effectively making them public
•Why it happens: Developers mechanically follow 'use getters/setters' without understanding why
•The problem: No real encapsulation—business logic scatters across the codebase
•The fix: Remove setters; add behavior methods. order.addItem(item) instead of order.setItems(list)

Anemic Domain Model Anti-Pattern
Java
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
// ❌ ANEMIC: Getters and setters for everything defeats encapsulation
public class Order {
    private List<OrderItem> items;
    private BigDecimal total;
    private OrderStatus status;
    
    public List<OrderItem> getItems() { return items; }
    public void setItems(List<OrderItem> items) { this.items = items; }
    
    public BigDecimal getTotal() { return total; }
    public void setTotal(BigDecimal total) { this.total = total; }
    
    public OrderStatus getStatus() { return status; }
    public void setStatus(OrderStatus status) { this.status = status; }
}
 
// Client code must manage all the logic:
order.getItems().add(newItem);
order.setTotal(order.getTotal().add(newItem.getPrice()));
if (order.getItems().size() > 10) {
    order.setStatus(OrderStatus.BULK);
}
// Business rules scattered everywhere, easily violated!
 
// ✅ RICH: Behavior methods encapsulate business logic
public class Order {
    private List<OrderItem> items = new ArrayList<>();
    private BigDecimal total = BigDecimal.ZERO;
    private OrderStatus status = OrderStatus.OPEN;
    
    public void addItem(OrderItem item) {
        items.add(item);
        total = total.add(item.getPrice());
        updateStatusBasedOnItems();
    }
    
    public void removeItem(OrderItem item) {
        if (items.remove(item)) {
            total = total.subtract(item.getPrice());
            updateStatusBasedOnItems();
        }
    }
    
    public BigDecimal getTotal() { return total; }  // Read-only
    public OrderStatus getStatus() { return status; }  // Read-only
    public List<OrderItem> getItems() {
        return Collections.unmodifiableList(items);  // Defensive!
    }
    
    private void updateStatusBasedOnItems() {
        status = items.size() > 10 ? OrderStatus.BULK : OrderStatus.OPEN;
    }
}

Anti-Pattern 2: Leaking Internal Collections

•What it is: Returning mutable references to internal collections
•Why it happens: Seems harmless—'I'm just giving them a list'
•The problem: Clients can modify internals, bypassing all your careful checks
•The fix: Return unmodifiable views or defensive copies

Anti-Pattern 3: Package-Private by Default

•What it is: Using package-private (default) visibility for fields out of laziness or habit
•Why it happens: Private feels 'restrictive'; package-private feels like a compromise
•The problem: Any class in the package can modify your internals—thousands of potential access points
•The fix: Always be explicit. Private is the default for fields; public is for interface methods

The Test of True Encapsulation

Ask yourself: 'If I completely changed how this class stores its data—different data types, different structure—would the public interface remain identical?' If yes, you've achieved encapsulation. If no, implementation details have leaked into your interface.

The Coupling Connection

Private fields and public interfaces aren't just about protecting invariants—they're fundamentally about managing coupling.

Coupling measures how much one module depends on another. High coupling means changes ripple widely; low coupling means changes are contained. The goal of good software design is low coupling between modules—achieved precisely through private fields and clean interfaces.

The Coupling Spectrum:

When client code depends on your class, the nature of that dependency determines how tightly coupled they are:

Content Coupling (Worst) — Client directly modifies your internal fields or implementation
Structural Coupling — Client knows your internal data structure
Data Coupling — Client knows the types you use internally
Behavioral Coupling (Best) — Client only knows what behaviors you offer

Private fields with behavioral public interfaces push you toward behavioral coupling—the loosest, most maintainable form.

Coupling Comparison
Java
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
// ❌ HIGH COUPLING: Client knows internal structure
public class UserRepository {
    public HashMap<String, User> userMap;  // Internal structure exposed!
}
 
// Client code:
User user = userRepository.userMap.get("user123");  // Knows it's a HashMap
userRepository.userMap.put("user456", newUser);     // Bypasses any logic
// If we change to TreeMap, database, or cache—client code breaks!
 
// ✅ LOW COUPLING: Client only knows interface
public class UserRepository {
    private final Map<String, User> users = new HashMap<>();
    
    public Optional<User> findById(String id) {
        return Optional.ofNullable(users.get(id));
    }
    
    public void save(User user) {
        validateUser(user);
        users.put(user.getId(), user);
        logAuditEvent(user);
    }
}
 
// Client code:
Optional<User> user = userRepository.findById("user123");
userRepository.save(newUser);
// Client doesn't know or care about HashMap. We can:
// - Change to TreeMap
// - Add caching layer
// - Move to database
// - Add sharding
// ALL WITHOUT TOUCHING CLIENT CODE!

The Dependency Direction Rule:

Coupling flows in the direction of visibility. When class A has public members, all clients become coupled to those members. Reducing public surface area reduces the potential for coupling.

This is why minimizing public interfaces isn't just aesthetic preference—it's engineering discipline. Every public member is a potential coupling point. The more coupling points, the more fragile your system becomes.

The Ripple Effect

In a system with 100 classes, a single public field that 50 classes depend on means any change to that field potentially affects 50 locations. With private fields and behavior methods, changes to internal representation affect only one location: the class itself.

Implementation Freedom in Practice

The ultimate benefit of private fields is implementation freedom—the ability to change how your class works internally without affecting any external code. This isn't theoretical; it's a practical capability that you'll exercise constantly throughout your career.

Real-World Scenario:

Imagine you've built a ShoppingCart class with private fields. Over time, requirements evolve:

Evolution of a Well-Encapsulated Class
Version	Internal Change	Public Interface
V1	Simple `ArrayList<CartItem>` in memory	Unchanged: `addItem()`, `removeItem()`, `getTotal()`
V2	Add HashMap for O(1) item lookup	Unchanged
V3	Persist to database for abandonment recovery	Unchanged
V4	Add Redis cache for performance	Unchanged
V5	Event sourcing for audit trail	Unchanged
V6	Sharded storage for scale	Unchanged

Six major architectural changes. Zero client code modifications.

Every client using cart.addItem(product, quantity) continues to work identically through all these evolutions. They don't know—and don't need to know—that their cart is now distributed across three Redis shards behind a database with event sourcing.

This is the power of encapsulation. It's not about hiding things to be secretive; it's about reserving the right to evolve.

The Alternative:

If the items field had been public, V2 would have broken all clients using the ArrayList API. V3 would have broken everyone. By V6, you'd have rewritten the entire system—or, more likely, you'd have caved and kept the legacy ArrayList forever, unable to scale.

The Freedom to Improve

Private fields give you the freedom to improve your implementation based on real-world feedback. You can profile, identify bottlenecks, optimize, refactor—all without coordinating releases with every client. This is how teams maintain velocity over years instead of grinding to a halt.

Summary: Private Fields, Public Interface

We've covered the foundational principle of data hiding—private fields with public interfaces. Let's consolidate the essential takeaways:

Key Takeaways

•Private is the default — Make all fields private unless there's a compelling reason otherwise
•Public interface is a contract — Every public member is a commitment you'll maintain forever
•Design behavior, not data access — Public methods should express what clients can do, not what data you hold
•Abstraction protects evolution — High-abstraction interfaces remain stable through implementation changes
•Coupling follows visibility — More public members means more coupling means more fragility
•Implementation freedom is the goal — Private fields let you change how without changing what

What's Next:

Now that we understand the principle of private fields and public interfaces, we'll explore how to control access to state—the practical mechanics of how clients read and modify object state through well-designed accessor patterns. You'll learn when getters and setters are appropriate, when they violate encapsulation, and how to design access mechanisms that protect your invariants while serving client needs.

Page Complete

You now understand the foundational principle of data hiding: private fields with public interfaces. This principle isn't merely syntactic—it's the mechanism by which well-designed classes protect their invariants, minimize coupling, and preserve the freedom to evolve. Apply this principle consistently, and your code will be dramatically more maintainable.

1 / 4

Loading learning content...

System Design (HLD)Encapsulation — Protecting Data & Behavior

Data Hiding Principles

LevelBeginner

Duration60 mins

TopicEncapsulation — Protecting Data & Behavior

1 / 4

Private Fields, Public Interface

The Invisible Barrier That Makes Software Work

What You Will Master

The Fundamental Principle

The principle can be stated simply:

Make all fields private. Expose only what clients genuinely need through carefully designed public methods.

Private Fields Enable

•Implementation Freedom — Change internal representation without affecting clients
•Invariant Protection — Guarantee that internal state remains valid
•Reduced Coupling — Clients depend only on behavior, not structure
•Controlled Evolution — Add features without breaking existing code
•Simplified Reasoning — Understand state changes through explicit methods

Public Fields Cause

•Implementation Lock-in — Any change breaks client code
•Invariant Violations — No way to enforce valid state
•Tight Coupling — Clients depend on internal structure
•Fragile Evolution — Every change risks cascading failures
•Scattered Logic — Validation and rules spread everywhere

Anatomy of a Well-Designed Class

Let's examine what a properly encapsulated class looks like by contrasting two approaches to the same problem: representing a bank account with balance constraints.

The Problem: Design a class representing a bank account that:

Has a balance that cannot go negative
Tracks the account holder's name
Records when the account was opened

BankAccount — Poor Design (Public Fields)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
// ❌ POOR DESIGN: Public fields expose implementation details
public class BankAccount {
    public double balance;           // Anyone can modify!
    public String holderName;        // Anyone can change!
    public Date openedDate;          // Anyone can alter history!
    
    public BankAccount(String holderName, double initialBalance) {
        this.holderName = holderName;
        this.balance = initialBalance;
        this.openedDate = new Date();
    }
}
 
// Client code (anywhere in the system):
BankAccount account = new BankAccount("Alice", 1000.0);
account.balance = -500.0;           // Negative balance! No protection.
account.holderName = "";            // Empty name! No validation.
account.openedDate = new Date(0);   // Altered history! No integrity.
 
// The class cannot protect its own invariants.
// Business rules must be enforced everywhere this class is used.

Now contrast this with a properly encapsulated design:

BankAccount — Good Design (Private Fields)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
// ✅ GOOD DESIGN: Private fields with controlled access
public class BankAccount {
    private double balance;
    private String holderName;
    private final Date openedDate;  // Immutable after construction
    
    public BankAccount(String holderName, double initialBalance) {
        validateHolderName(holderName);
        validateBalance(initialBalance);
        
        this.holderName = holderName;
        this.balance = initialBalance;
        this.openedDate = new Date();
    }
    
    // ===== PUBLIC INTERFACE: What clients can do =====
    
    public double getBalance() {
        return balance;
    }
    
    public String getHolderName() {
        return holderName;
    }
    
    public Date getOpenedDate() {
        return new Date(openedDate.getTime());  // Defensive copy!
    }
    
    public void deposit(double amount) {
        if (amount <= 0) {
            throw new IllegalArgumentException("Deposit must be positive");
        }
        balance += amount;
    }
    
    public void withdraw(double amount) {
        if (amount <= 0) {
            throw new IllegalArgumentException("Withdrawal must be positive");
        }
        if (amount > balance) {
            throw new IllegalStateException("Insufficient funds");
        }
        balance -= amount;
    }
    
    // ===== PRIVATE IMPLEMENTATION: How invariants are enforced =====
    
    private void validateHolderName(String name) {
        if (name == null || name.trim().isEmpty()) {
            throw new IllegalArgumentException("Holder name required");
        }
    }
    
    private void validateBalance(double amount) {
        if (amount < 0) {
            throw new IllegalArgumentException("Initial balance cannot be negative");
        }
    }
}
 
// Client code is now FORCED to go through the interface:
BankAccount account = new BankAccount("Alice", 1000.0);
account.deposit(500.0);    // ✓ Valid operation
account.withdraw(200.0);   // ✓ Valid operation
account.withdraw(5000.0);  // ✗ Throws exception - invariant protected!

The Critical Difference

The Public Interface as Contract

The Contract Promises:

Existence — The public member will continue to exist in future versions
Signature Stability — The parameters and return type won't change incompatibly
Behavioral Consistency — The method will continue to do what it currently does
Semantic Integrity — The meaning of the member won't fundamentally change

Contract Implications of Public vs Private
Aspect	Public Member	Private Member
Change Freedom	Nearly frozen—changes break clients	Complete freedom to refactor
Documentation Burden	Must be documented for external use	Internal comments sufficient
Testing Requirements	Must test for backward compatibility	Test for correctness only
Version Coupling	Clients lock to your version	Clients unaffected by changes
Design Pressure	Must be near-perfect upfront	Can iterate and improve freely

The asymmetry is stark: Making something private gives you complete freedom; making something public chains you to that decision forever.

The Minimum Interface Principle:

Expose the smallest public interface that enables clients to accomplish their goals. Every additional public member is additional maintenance burden and additional constraint on your future evolution.

Ask yourself for each potential public member: "Does the client need this, or is it just convenient?" Convenience is not sufficient justification for public exposure.

The Ratchet Effect

Designing the Public Interface

Characteristics of Excellent Public Interfaces

•Behavior-Centric — Express what clients can do, not what data you hold. account.deposit(amount) rather than account.balance += amount.
•Self-Documenting — Method names clearly communicate purpose and intent. A reader shouldn't need to inspect implementation to understand usage.
•Minimal — Include only what clients genuinely need. Resist the urge to expose 'just in case' features.
•Consistent — Follow naming conventions and patterns uniformly. If one method is getUserById(), don't have another called fetchOrderForId().
•Complete — Enable all legitimate use cases without requiring workarounds. If clients can't accomplish their goals, they'll resort to hacks.
•Orthogonal — Each method does one thing. Avoid overlapping functionality that can be composed from other public methods.

The Abstraction Level Rule:

Your public interface should operate at a higher abstraction level than your private implementation. This is what allows the interface to remain stable while the implementation evolves.

Consider the difference:

Abstraction Levels in Interface Design
Low Abstraction (Poor)	High Abstraction (Good)
`setBalanceField(double value)`	`deposit(double amount)`
`getInternalHashMap()`	`findUserById(String id)`
`sqlQuery(String query)`	`findActiveOrders()`
`setDatabaseConnection(Connection c)`	`connect(String url)`
`getArrayList()`	`getItems()` returning List interface

The low-abstraction versions expose implementation details: field names, internal data structures, database specifics. Changing any of these requires changing the interface.

The Shopping Cart Test

Common Anti-Patterns in Field Visibility

Anti-Pattern 1: The Anemic Domain Model

•What it is: Classes with all private fields but trivial getters and setters for everything, effectively making them public
•Why it happens: Developers mechanically follow 'use getters/setters' without understanding why
•The problem: No real encapsulation—business logic scatters across the codebase
•The fix: Remove setters; add behavior methods. order.addItem(item) instead of order.setItems(list)

Anemic Domain Model Anti-Pattern
Java
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
// ❌ ANEMIC: Getters and setters for everything defeats encapsulation
public class Order {
    private List<OrderItem> items;
    private BigDecimal total;
    private OrderStatus status;
    
    public List<OrderItem> getItems() { return items; }
    public void setItems(List<OrderItem> items) { this.items = items; }
    
    public BigDecimal getTotal() { return total; }
    public void setTotal(BigDecimal total) { this.total = total; }
    
    public OrderStatus getStatus() { return status; }
    public void setStatus(OrderStatus status) { this.status = status; }
}
 
// Client code must manage all the logic:
order.getItems().add(newItem);
order.setTotal(order.getTotal().add(newItem.getPrice()));
if (order.getItems().size() > 10) {
    order.setStatus(OrderStatus.BULK);
}
// Business rules scattered everywhere, easily violated!
 
// ✅ RICH: Behavior methods encapsulate business logic
public class Order {
    private List<OrderItem> items = new ArrayList<>();
    private BigDecimal total = BigDecimal.ZERO;
    private OrderStatus status = OrderStatus.OPEN;
    
    public void addItem(OrderItem item) {
        items.add(item);
        total = total.add(item.getPrice());
        updateStatusBasedOnItems();
    }
    
    public void removeItem(OrderItem item) {
        if (items.remove(item)) {
            total = total.subtract(item.getPrice());
            updateStatusBasedOnItems();
        }
    }
    
    public BigDecimal getTotal() { return total; }  // Read-only
    public OrderStatus getStatus() { return status; }  // Read-only
    public List<OrderItem> getItems() {
        return Collections.unmodifiableList(items);  // Defensive!
    }
    
    private void updateStatusBasedOnItems() {
        status = items.size() > 10 ? OrderStatus.BULK : OrderStatus.OPEN;
    }
}

Anti-Pattern 2: Leaking Internal Collections

•What it is: Returning mutable references to internal collections
•Why it happens: Seems harmless—'I'm just giving them a list'
•The problem: Clients can modify internals, bypassing all your careful checks
•The fix: Return unmodifiable views or defensive copies

Anti-Pattern 3: Package-Private by Default

•What it is: Using package-private (default) visibility for fields out of laziness or habit
•Why it happens: Private feels 'restrictive'; package-private feels like a compromise
•The problem: Any class in the package can modify your internals—thousands of potential access points
•The fix: Always be explicit. Private is the default for fields; public is for interface methods

The Test of True Encapsulation

The Coupling Connection

Private fields and public interfaces aren't just about protecting invariants—they're fundamentally about managing coupling.

The Coupling Spectrum:

When client code depends on your class, the nature of that dependency determines how tightly coupled they are:

Content Coupling (Worst) — Client directly modifies your internal fields or implementation
Structural Coupling — Client knows your internal data structure
Data Coupling — Client knows the types you use internally
Behavioral Coupling (Best) — Client only knows what behaviors you offer

Private fields with behavioral public interfaces push you toward behavioral coupling—the loosest, most maintainable form.

Coupling Comparison
Java
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
// ❌ HIGH COUPLING: Client knows internal structure
public class UserRepository {
    public HashMap<String, User> userMap;  // Internal structure exposed!
}
 
// Client code:
User user = userRepository.userMap.get("user123");  // Knows it's a HashMap
userRepository.userMap.put("user456", newUser);     // Bypasses any logic
// If we change to TreeMap, database, or cache—client code breaks!
 
// ✅ LOW COUPLING: Client only knows interface
public class UserRepository {
    private final Map<String, User> users = new HashMap<>();
    
    public Optional<User> findById(String id) {
        return Optional.ofNullable(users.get(id));
    }
    
    public void save(User user) {
        validateUser(user);
        users.put(user.getId(), user);
        logAuditEvent(user);
    }
}
 
// Client code:
Optional<User> user = userRepository.findById("user123");
userRepository.save(newUser);
// Client doesn't know or care about HashMap. We can:
// - Change to TreeMap
// - Add caching layer
// - Move to database
// - Add sharding
// ALL WITHOUT TOUCHING CLIENT CODE!

The Dependency Direction Rule:

Coupling flows in the direction of visibility. When class A has public members, all clients become coupled to those members. Reducing public surface area reduces the potential for coupling.

The Ripple Effect

Implementation Freedom in Practice

Real-World Scenario:

Imagine you've built a ShoppingCart class with private fields. Over time, requirements evolve:

Evolution of a Well-Encapsulated Class
Version	Internal Change	Public Interface
V1	Simple `ArrayList<CartItem>` in memory	Unchanged: `addItem()`, `removeItem()`, `getTotal()`
V2	Add HashMap for O(1) item lookup	Unchanged
V3	Persist to database for abandonment recovery	Unchanged
V4	Add Redis cache for performance	Unchanged
V5	Event sourcing for audit trail	Unchanged
V6	Sharded storage for scale	Unchanged

Six major architectural changes. Zero client code modifications.

This is the power of encapsulation. It's not about hiding things to be secretive; it's about reserving the right to evolve.

The Alternative:

The Freedom to Improve

Summary: Private Fields, Public Interface

We've covered the foundational principle of data hiding—private fields with public interfaces. Let's consolidate the essential takeaways:

Key Takeaways

•Private is the default — Make all fields private unless there's a compelling reason otherwise
•Public interface is a contract — Every public member is a commitment you'll maintain forever
•Design behavior, not data access — Public methods should express what clients can do, not what data you hold
•Abstraction protects evolution — High-abstraction interfaces remain stable through implementation changes
•Coupling follows visibility — More public members means more coupling means more fragility
•Implementation freedom is the goal — Private fields let you change how without changing what

What's Next:

Page Complete

1 / 4