System Design (LLD)Liskov Substitution Principle

Invariants in Inheritance

LevelIntermediate

Duration75 mins

TopicLiskov Substitution Principle

1 / 4

Class Invariants - Always-True Conditions

The Invisible Contract

Imagine you're working with a BankAccount class in a financial system. You call getBalance() and receive -$500,000. But wait—wasn't this account type supposed to be a "Non-Negative Balance Account"? How did this happen?

The answer lies in class invariants—properties that must always be true for an object, from the moment it's created until the moment it's destroyed. When these invariants are violated, objects enter impossible states, and systems fail in unpredictable ways.

An invariant isn't just documentation. It's a contract that every method must maintain.

The BankAccount that went negative had an invariant: balance >= 0. Somewhere, some method—perhaps in a subclass—violated this invariant. The result? Data corruption flowing through the system, affecting calculations, reports, and potentially real-world financial decisions.

This page explores the concept of class invariants: what they are, why they're critical to reliable software, and how they form the foundation of the Liskov Substitution Principle's behavioral contracts.

What You Will Learn

By the end of this page, you will understand what class invariants are, why they're essential for object integrity, how to identify invariants in your designs, and how invariants create implicit contracts that every method must honor. You'll learn to think about objects not just in terms of behavior, but in terms of the guarantees they must maintain.

What Are Class Invariants?

A class invariant is a condition that is always true for every valid instance of a class, throughout its entire lifetime. From the moment an object is constructed until the moment it goes out of scope, invariants must hold.

Formal definition:

A class invariant is a logical condition involving the instance variables of a class that must be true before and after every method call. If the invariant ever becomes false, the object is in an invalid state.

Let's unpack this carefully:

"Always true" means there's never a moment (between method calls) when the invariant can be false
"Before and after every method" means methods can temporarily break invariants during execution, but must restore them before returning
"Invalid state" means the object's data no longer represents a meaningful entity

Invariants are different from preconditions and postconditions:

Preconditions: What must be true before a method is called
Postconditions: What must be true after a method completes
Invariants: What must be true always, independent of any specific method

Invariants vs Preconditions vs Postconditions
Concept	Scope	Who Ensures It	When It Must Hold
Precondition	Single method call	The caller	Before the method begins execution
Postcondition	Single method call	The method implementation	After the method completes execution
Invariant	Object lifetime	Every method in the class	Always, except during method execution

invariant-basics
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
67
68
69
/**
 * A DateRange class with a clear invariant:
 * INVARIANT: startDate <= endDate (start never after end)
 * 
 * This invariant must be true:
 * - After construction
 * - After every method call
 * - During the object's entire lifetime
 */
class DateRange {
    private startDate: Date;
    private endDate: Date;
    
    // INVARIANT: startDate <= endDate
    
    constructor(start: Date, end: Date) {
        // Check invariant at construction time
        if (start > end) {
            throw new Error("Invalid DateRange: start cannot be after end");
        }
        this.startDate = start;
        this.endDate = end;
        // Invariant established: startDate <= endDate ✓
    }
    
    /**
     * Move the start date earlier.
     * This always maintains the invariant since we're moving start earlier.
     */
    extendStart(newStart: Date): void {
        if (newStart > this.endDate) {
            throw new Error("Cannot set start after end date");
        }
        this.startDate = newStart;
        // Invariant maintained: startDate <= endDate ✓
    }
    
    /**
     * Move the end date later.
     * This always maintains the invariant since we're moving end later.
     */
    extendEnd(newEnd: Date): void {
        if (newEnd < this.startDate) {
            throw new Error("Cannot set end before start date");
        }
        this.endDate = newEnd;
        // Invariant maintained: startDate <= endDate ✓
    }
    
    /**
     * Shift the entire range by a number of days.
     * Invariant is maintained because both dates move equally.
     */
    shiftDays(days: number): void {
        const msPerDay = 24 * 60 * 60 * 1000;
        const shift = days * msPerDay;
        
        // Both dates shift together - invariant cannot be violated
        this.startDate = new Date(this.startDate.getTime() + shift);
        this.endDate = new Date(this.endDate.getTime() + shift);
        // Invariant maintained: startDate <= endDate ✓
    }
    
    getDurationDays(): number {
        const ms = this.endDate.getTime() - this.startDate.getTime();
        return Math.floor(ms / (24 * 60 * 60 * 1000));
        // Note: This method relies on the invariant to return non-negative
    }
}

The Invariant-Method Relationship

Notice how every method in DateRange is designed with the invariant in mind. Methods either verify that their changes won't violate the invariant (and throw if they would), or they make changes that structurally cannot violate the invariant (like shifting both dates together). This is the discipline invariant thinking requires.

Types of Class Invariants

Invariants come in various forms, from simple value constraints to complex relationships between multiple fields. Understanding these categories helps you identify and document invariants in your own classes.

1. Value Constraints (Bounds, Ranges, Allowed Values)

These invariants restrict individual fields to specific domains:

value-constraints
TypeScript
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
class Percentage {
    private value: number;
    
    // INVARIANT: 0 <= value <= 100
    
    constructor(value: number) {
        if (value < 0 || value > 100) {
            throw new Error("Percentage must be between 0 and 100");
        }
        this.value = value;
    }
    
    add(other: Percentage): Percentage {
        // Result capped to maintain invariant
        const sum = Math.min(100, this.value + other.value);
        return new Percentage(sum);
    }
}
 
class EmailAddress {
    private readonly address: string;
    
    // INVARIANT: address is a valid email format (contains @ and domain)
    
    constructor(address: string) {
        if (!this.isValidEmail(address)) {
            throw new Error("Invalid email address format");
        }
        this.address = address;
    }
    
    private isValidEmail(email: string): boolean {
        // Simplified validation
        const pattern = /^[^\s@]+@[^\s@]+\.[^\s@]+$/;
        return pattern.test(email);
    }
}
 
class AgeBracket {
    private minAge: number;
    private maxAge: number;
    
    // INVARIANT: 0 <= minAge <= maxAge <= 150
    
    constructor(minAge: number, maxAge: number) {
        if (minAge < 0 || maxAge > 150 || minAge > maxAge) {
            throw new Error("Invalid age bracket");
        }
        this.minAge = minAge;
        this.maxAge = maxAge;
    }
}

2. Relationship Invariants (Between Fields)

These invariants express that multiple fields must maintain a specific relationship:

relationship-invariants
TypeScript
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
class Rectangle {
    private width: number;
    private height: number;
    private area: number;  // Cached for performance
    
    // INVARIANT: area === width * height (always synchronized)
    
    constructor(width: number, height: number) {
        this.width = width;
        this.height = height;
        this.area = width * height;  // Establish invariant
    }
    
    setWidth(newWidth: number): void {
        this.width = newWidth;
        this.area = this.width * this.height;  // Maintain invariant
    }
    
    setHeight(newHeight: number): void {
        this.height = newHeight;
        this.area = this.width * this.height;  // Maintain invariant
    }
    
    getArea(): number {
        // Relies on invariant - no recalculation needed
        return this.area;
    }
}
 
class SortedList<T> {
    private items: T[];
    private comparator: (a: T, b: T) => number;
    
    // INVARIANT: For all i < j: comparator(items[i], items[j]) <= 0
    // (Items are always in sorted order)
    
    constructor(comparator: (a: T, b: T) => number) {
        this.items = [];
        this.comparator = comparator;
    }
    
    add(item: T): void {
        // Binary search to find insertion point - maintains sort invariant
        let left = 0, right = this.items.length;
        while (left < right) {
            const mid = Math.floor((left + right) / 2);
            if (this.comparator(this.items[mid], item) <= 0) {
                left = mid + 1;
            } else {
                right = mid;
            }
        }
        this.items.splice(left, 0, item);
        // Invariant maintained: items remain sorted
    }
    
    getMin(): T | undefined {
        // Relies on invariant: smallest is always at index 0
        return this.items[0];
    }
    
    getMax(): T | undefined {
        // Relies on invariant: largest is always at last index
        return this.items[this.items.length - 1];
    }
}

3. Structural Invariants (Data Structure Integrity)

These invariants ensure that internal data structures remain valid:

structural-invariants
TypeScript
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
67
68
69
class BinarySearchTree<T> {
    private root: TreeNode<T> | null = null;
    private comparator: (a: T, b: T) => number;
    
    // INVARIANT: For every node N:
    //   - All values in N.left subtree < N.value
    //   - All values in N.right subtree > N.value
    // (Binary Search Tree property)
    
    constructor(comparator: (a: T, b: T) => number) {
        this.comparator = comparator;
    }
    
    insert(value: T): void {
        // Insert maintains BST invariant by choosing correct position
        if (!this.root) {
            this.root = { value, left: null, right: null };
            return;
        }
        
        this.insertRecursive(this.root, value);
        // Invariant maintained: BST property preserved
    }
    
    private insertRecursive(node: TreeNode<T>, value: T): void {
        const comparison = this.comparator(value, node.value);
        
        if (comparison < 0) {
            // Value is less - goes in left subtree (maintains invariant)
            if (node.left) {
                this.insertRecursive(node.left, value);
            } else {
                node.left = { value, left: null, right: null };
            }
        } else {
            // Value is greater/equal - goes in right subtree (maintains invariant)
            if (node.right) {
                this.insertRecursive(node.right, value);
            } else {
                node.right = { value, left: null, right: null };
            }
        }
    }
    
    contains(value: T): boolean {
        // Search relies on BST invariant for O(log n) lookup
        return this.searchRecursive(this.root, value);
    }
    
    private searchRecursive(node: TreeNode<T> | null, value: T): boolean {
        if (!node) return false;
        
        const comparison = this.comparator(value, node.value);
        if (comparison === 0) return true;
        
        // Invariant guarantees we only need to search one subtree
        if (comparison < 0) {
            return this.searchRecursive(node.left, value);
        } else {
            return this.searchRecursive(node.right, value);
        }
    }
}
 
interface TreeNode<T> {
    value: T;
    left: TreeNode<T> | null;
    right: TreeNode<T> | null;
}

4. Count/Size Invariants (Synchronized Counts)

These invariants ensure that size or count fields accurately reflect the actual data:

count-invariants
TypeScript
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
class LinkedList<T> {
    private head: ListNode<T> | null = null;
    private tail: ListNode<T> | null = null;
    private count: number = 0;
    
    // INVARIANT 1: count === actual number of nodes in list
    // INVARIANT 2: if count === 0, then head === null && tail === null
    // INVARIANT 3: if count > 0, then head !== null && tail !== null
    // INVARIANT 4: tail.next === null (tail is always the last node)
    
    add(value: T): void {
        const newNode: ListNode<T> = { value, next: null };
        
        if (this.tail) {
            this.tail.next = newNode;
            this.tail = newNode;
        } else {
            this.head = newNode;
            this.tail = newNode;
        }
        
        this.count++;  // Maintain count invariant
        // All invariants maintained
    }
    
    removeFirst(): T | undefined {
        if (!this.head) return undefined;
        
        const value = this.head.value;
        this.head = this.head.next;
        
        if (!this.head) {
            this.tail = null;  // Maintain invariant 2
        }
        
        this.count--;  // Maintain count invariant
        // All invariants maintained
        
        return value;
    }
    
    size(): number {
        // Relies on count invariant - O(1) instead of traversing
        return this.count;
    }
    
    isEmpty(): boolean {
        // Relies on invariant 2
        return this.count === 0;
    }
}
 
interface ListNode<T> {
    value: T;
    next: ListNode<T> | null;
}

Documenting Invariants

The most important step in working with invariants is making them explicit. Document invariants as comments in your class definition. Future maintainers (including future you) must understand what properties the class guarantees, so they don't accidentally break them when adding new methods.

Why Invariants Matter

Invariants might seem like additional overhead—extra checks, extra thinking, extra documentation. But they provide profound benefits that compound as systems grow in size and complexity.

1. Trust and Simplification

When invariants are guaranteed, code can make assumptions without defensive checks:

Without Invariant Guarantees

•Every method must validate inputs defensively
•Callers can't trust returned values
•Business logic drowns in validation code
•Bugs hide in forgotten edge cases
•Code becomes verbose and brittle

With Invariant Guarantees

•Methods assume valid state on entry
•Returned objects are always valid
•Business logic is clear and focused
•Invalid states are prevented by design
•Code is concise and robust

trust-simplification
TypeScript
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
// WITHOUT invariant guarantees - defensive everywhere
class UntrustedRectangle {
    width: number | null = null;
    height: number | null = null;
    
    getArea(): number | null {
        // Must check everything because nothing is guaranteed
        if (this.width === null || this.height === null) {
            return null;
        }
        if (this.width < 0 || this.height < 0) {
            return null;  // Or throw? Caller doesn't know what to expect
        }
        return this.width * this.height;
    }
    
    getPerimeter(): number | null {
        // Same defensive checks repeated everywhere
        if (this.width === null || this.height === null) {
            return null;
        }
        if (this.width < 0 || this.height < 0) {
            return null;
        }
        return 2 * (this.width + this.height);
    }
}
 
// WITH invariant guarantees - clean and focused
class TrustedRectangle {
    // INVARIANT: width > 0, height > 0 (always positive, never null)
    
    private constructor(
        private readonly width: number,
        private readonly height: number
    ) {}
    
    static create(width: number, height: number): TrustedRectangle {
        // Invariant enforced at construction
        if (width <= 0 || height <= 0) {
            throw new Error("Dimensions must be positive");
        }
        return new TrustedRectangle(width, height);
    }
    
    getArea(): number {
        // No checks needed - invariant guarantees valid dimensions
        return this.width * this.height;
    }
    
    getPerimeter(): number {
        // Clean implementation relying on invariant
        return 2 * (this.width + this.height);
    }
    
    scale(factor: number): TrustedRectangle {
        if (factor <= 0) {
            throw new Error("Scale factor must be positive");
        }
        // Result automatically satisfies invariant
        return new TrustedRectangle(this.width * factor, this.height * factor);
    }
}

2. Reasoning About Correctness

Invariants make it possible to reason about program correctness systematically:

During code review: Verify each method maintains invariants
During debugging: If invariant is violated, you know exactly what broke
During testing: Tests can verify invariants hold after each operation
During design: Invariants guide API design and reveal impossible operations

3. Performance Optimization

Invariants enable optimizations that would otherwise be unsafe:

invariant-optimization
TypeScript
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
67
68
69
class SortedArray<T> {
    private items: T[];
    private comparator: (a: T, b: T) => number;
    
    // INVARIANT: items is always sorted according to comparator
    
    constructor(comparator: (a: T, b: T) => number) {
        this.items = [];
        this.comparator = comparator;
    }
    
    /**
     * Because of the invariant, we can use binary search.
     * O(log n) instead of O(n) for unsorted arrays.
     */
    contains(item: T): boolean {
        return this.binarySearch(item) !== -1;
    }
    
    /**
     * Find minimum in O(1) because invariant guarantees smallest is first.
     */
    min(): T | undefined {
        return this.items[0];  // Invariant: first is always smallest
    }
    
    /**
     * Find maximum in O(1) because invariant guarantees largest is last.
     */
    max(): T | undefined {
        return this.items[this.items.length - 1];  // Invariant: last is always largest
    }
    
    /**
     * Merge two sorted arrays in O(n + m) because both maintain sort invariant.
     * Without the invariant, this would require O((n+m) log(n+m)) to sort.
     */
    merge(other: SortedArray<T>): SortedArray<T> {
        const result = new SortedArray<T>(this.comparator);
        let i = 0, j = 0;
        
        // Merge in sorted order - only possible because both are sorted
        while (i < this.items.length && j < other.items.length) {
            if (this.comparator(this.items[i], other.items[j]) <= 0) {
                result.items.push(this.items[i++]);
            } else {
                result.items.push(other.items[j++]);
            }
        }
        
        // Append remaining items (already sorted)
        while (i < this.items.length) result.items.push(this.items[i++]);
        while (j < other.items.length) result.items.push(other.items[j++]);
        
        return result;  // Result also maintains invariant
    }
    
    private binarySearch(item: T): number {
        let left = 0, right = this.items.length - 1;
        while (left <= right) {
            const mid = Math.floor((left + right) / 2);
            const cmp = this.comparator(this.items[mid], item);
            if (cmp === 0) return mid;
            if (cmp < 0) left = mid + 1;
            else right = mid - 1;
        }
        return -1;
    }
}

Breaking Invariants Breaks Optimizations

If any code path allows the SortedArray to become unsorted, every optimization based on the sort invariant fails. Binary search returns wrong results. Min/max become incorrect. Merge produces unsorted output. This is why invariant violations are so dangerous—they silently corrupt behavior across the entire class.

Invariants as Implicit Contracts

Here's a critical insight: Class invariants create implicit contracts between the class and all code that uses it.

When you define a class with invariants, you're making promises to callers:

"This BankAccount will never have a negative balance"
"This DateRange will always have start ≤ end"
"This SortedList will always be in order"

Callers rely on these promises. They write code that assumes the invariants are true. If you violate the invariants, you break that code—often in subtle, hard-to-detect ways.

implicit-contracts
TypeScript
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
class NonNegativeBalance {
    // INVARIANT: balance >= 0
    private balance: number;
    
    constructor(initialBalance: number) {
        if (initialBalance < 0) {
            throw new Error("Initial balance cannot be negative");
        }
        this.balance = initialBalance;
    }
    
    deposit(amount: number): void {
        if (amount < 0) {
            throw new Error("Cannot deposit negative amount");
        }
        this.balance += amount;
        // Invariant maintained: adding positive to non-negative stays non-negative
    }
    
    withdraw(amount: number): void {
        if (amount < 0) {
            throw new Error("Cannot withdraw negative amount");
        }
        if (amount > this.balance) {
            throw new Error("Insufficient funds");
        }
        this.balance -= amount;
        // Invariant maintained: we checked amount <= balance
    }
    
    getBalance(): number {
        return this.balance;  // Caller can trust this is >= 0
    }
}
 
// Caller code relies on the invariant
function calculateInterest(account: NonNegativeBalance): number {
    const balance = account.getBalance();
    
    // This code TRUSTS that balance >= 0
    // No check needed because the invariant guarantees it
    const rate = 0.05;
    return balance * rate;  // Always non-negative
}
 
function displayBalance(account: NonNegativeBalance): string {
    const balance = account.getBalance();
    
    // Safe to use without checking for negative
    // The invariant guarantees balance >= 0
    return `$${balance.toFixed(2)}`;
}
 
function transferSafely(
    from: NonNegativeBalance, 
    to: NonNegativeBalance, 
    amount: number
): void {
    // This operation is safe because both accounts maintain invariants
    from.withdraw(amount);  // Will throw if insufficient funds
    to.deposit(amount);
    // If we reach here, both accounts still satisfy invariants
}

The Trust Chain:

Invariants create a chain of trust throughout your codebase:

Class guarantees invariant → Methods ensure invariants hold after every operation
Methods rely on invariant → Internal implementation doesn't re-check what's guaranteed
Callers rely on invariant → External code trusts the class promises
Downstream relies on caller → Systems built on top trust their building blocks

If any link in this chain breaks—if any method fails to maintain the invariant—the corruption propagates through the entire chain. Code that was correct becomes incorrect, not because it changed, but because something it trusted broke its promise.

Make Invariants Unbreakable

The best way to maintain invariants is to make them impossible to break. Use private fields with controlled accessors. Validate at construction time. Make objects immutable where possible. The goal is to structure your class so that no sequence of method calls can ever create an invalid state.

Strategies for Maintaining Invariants
Strategy	How It Works	When to Use
Private fields only	External code cannot directly modify state	Always - default approach
Validate in constructor	Reject invalid initial states	Always - establish invariant at birth
Validate in mutators	Check before every state change	When mutation is necessary
Immutability	Objects can't change, so invariants can't be violated post-construction	When feasible - strongest guarantee
Factory methods	Control object creation to ensure valid states	Complex initialization logic
Defensive copies	Prevent callers from mutating internal state	When returning mutable internals

Identifying Invariants in Existing Code

How do you uncover the invariants in code that doesn't document them? This is a critical skill, especially when working with legacy systems or third-party libraries.

Discovery Techniques:

How to Find Hidden Invariants

•Look at constructors: What validation happens? What must be true for construction to succeed?
•Examine field types: Are they non-nullable? Bounded? What types were chosen?
•Read setter/mutator methods: What throws exceptions? What conditions are checked?
•Study getter methods: What assumptions do they make about internal state?
•Analyze assertions and checks: What conditions does the code verify?
•Check test cases: What properties do tests verify? What edge cases are covered?
•Review documentation: Even informal comments hint at expected properties

discovering-invariants
TypeScript
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
// Given this class, what are its invariants?
class Order {
    private items: OrderItem[];
    private status: OrderStatus;
    private total: number;
    private createdAt: Date;
    private shippedAt: Date | null;
    
    constructor(items: OrderItem[]) {
        if (items.length === 0) {
            throw new Error("Order must have at least one item");
        }
        this.items = [...items];  // Defensive copy
        this.status = OrderStatus.PENDING;
        this.total = this.calculateTotal();
        this.createdAt = new Date();
        this.shippedAt = null;
    }
    
    private calculateTotal(): number {
        return this.items.reduce((sum, item) => sum + item.price * item.quantity, 0);
    }
    
    ship(): void {
        if (this.status !== OrderStatus.PENDING) {
            throw new Error("Can only ship pending orders");
        }
        this.status = OrderStatus.SHIPPED;
        this.shippedAt = new Date();
    }
    
    getItems(): OrderItem[] {
        return [...this.items];  // Defensive copy - can't modify internal
    }
}
 
// DISCOVERED INVARIANTS:
// 1. items.length > 0 (checked in constructor)
// 2. total === sum of item prices * quantities (set at construction)
// 3. createdAt is never null (set at construction)
// 4. If status === SHIPPED, then shippedAt !== null
// 5. If status === PENDING, then shippedAt === null
// 6. items array is immutable after construction (defensive copies)
// 7. status transitions: PENDING -> SHIPPED (never backwards)

Common Invariant Patterns:

Common Invariants in Software Systems
Domain	Common Invariants
Financial	balance >= 0; credits - debits = balance; transaction amounts > 0
Collections	size >= 0; size = actual count; indices < size
Date/Time	start <= end; duration >= 0; times are in valid ranges
Graphs	edge connects valid nodes; no orphan edges; degrees are accurate
Trees	one root; no cycles; parent-child links consistent
Caches	capacity >= 0; size <= capacity; entries match stored keys
User Data	id is unique; email format is valid; required fields are non-null

Document What You Discover

When you discover undocumented invariants, add them to the code as comments. Future maintainers will thank you. Better yet, add assertions or validation that explicitly checks invariants, making violations fail fast and loudly.

Summary: The Foundation of Object Integrity

We've explored class invariants in depth. Let's consolidate the key insights:

Key Takeaways

•Invariants are always-true conditions — Properties that must hold from object creation through its entire lifetime
•Invariants come in many forms — Value constraints, field relationships, structural properties, synchronized counts
•Every method must maintain invariants — Methods can temporarily break invariants during execution, but must restore them before returning
•Invariants enable trust and optimization — Code can assume invariant properties without defensive checks
•Invariants create implicit contracts — Callers rely on invariant properties; violating them breaks downstream code
•Document your invariants explicitly — Future maintainers must understand what properties the class guarantees

What's Next:

Understanding invariants is crucial preparation for the next topic: Preserving Invariants in Subclasses. When inheritance enters the picture, invariants become even more critical—and more easily violated.

A subclass inherits not just the code of its parent, but also its invariants. If a subclass weakens or breaks these invariants, it violates the Liskov Substitution Principle, causing polymorphic code to fail in subtle, dangerous ways.

You'll learn exactly how subclasses can accidentally break parent invariants, and how to design inheritance hierarchies that maintain invariant integrity across the entire type hierarchy.

Page Complete

You now understand what class invariants are and why they're fundamental to object integrity. Invariants aren't just nice-to-have documentation—they're the contracts that make polymorphism safe and enable code to trust the objects it works with. Next, we'll explore how subclasses must preserve these invariants to maintain LSP compliance.

1 / 4

Loading learning content...

System Design (LLD)Liskov Substitution Principle

Invariants in Inheritance

LevelIntermediate

Duration75 mins

TopicLiskov Substitution Principle

1 / 4

Class Invariants - Always-True Conditions

The Invisible Contract

An invariant isn't just documentation. It's a contract that every method must maintain.

What You Will Learn

What Are Class Invariants?

Formal definition:

A class invariant is a logical condition involving the instance variables of a class that must be true before and after every method call. If the invariant ever becomes false, the object is in an invalid state.

Let's unpack this carefully:

"Always true" means there's never a moment (between method calls) when the invariant can be false
"Before and after every method" means methods can temporarily break invariants during execution, but must restore them before returning
"Invalid state" means the object's data no longer represents a meaningful entity

Invariants are different from preconditions and postconditions:

Preconditions: What must be true before a method is called
Postconditions: What must be true after a method completes
Invariants: What must be true always, independent of any specific method

Invariants vs Preconditions vs Postconditions
Concept	Scope	Who Ensures It	When It Must Hold
Precondition	Single method call	The caller	Before the method begins execution
Postcondition	Single method call	The method implementation	After the method completes execution
Invariant	Object lifetime	Every method in the class	Always, except during method execution

invariant-basics
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
67
68
69
/**
 * A DateRange class with a clear invariant:
 * INVARIANT: startDate <= endDate (start never after end)
 * 
 * This invariant must be true:
 * - After construction
 * - After every method call
 * - During the object's entire lifetime
 */
class DateRange {
    private startDate: Date;
    private endDate: Date;
    
    // INVARIANT: startDate <= endDate
    
    constructor(start: Date, end: Date) {
        // Check invariant at construction time
        if (start > end) {
            throw new Error("Invalid DateRange: start cannot be after end");
        }
        this.startDate = start;
        this.endDate = end;
        // Invariant established: startDate <= endDate ✓
    }
    
    /**
     * Move the start date earlier.
     * This always maintains the invariant since we're moving start earlier.
     */
    extendStart(newStart: Date): void {
        if (newStart > this.endDate) {
            throw new Error("Cannot set start after end date");
        }
        this.startDate = newStart;
        // Invariant maintained: startDate <= endDate ✓
    }
    
    /**
     * Move the end date later.
     * This always maintains the invariant since we're moving end later.
     */
    extendEnd(newEnd: Date): void {
        if (newEnd < this.startDate) {
            throw new Error("Cannot set end before start date");
        }
        this.endDate = newEnd;
        // Invariant maintained: startDate <= endDate ✓
    }
    
    /**
     * Shift the entire range by a number of days.
     * Invariant is maintained because both dates move equally.
     */
    shiftDays(days: number): void {
        const msPerDay = 24 * 60 * 60 * 1000;
        const shift = days * msPerDay;
        
        // Both dates shift together - invariant cannot be violated
        this.startDate = new Date(this.startDate.getTime() + shift);
        this.endDate = new Date(this.endDate.getTime() + shift);
        // Invariant maintained: startDate <= endDate ✓
    }
    
    getDurationDays(): number {
        const ms = this.endDate.getTime() - this.startDate.getTime();
        return Math.floor(ms / (24 * 60 * 60 * 1000));
        // Note: This method relies on the invariant to return non-negative
    }
}

The Invariant-Method Relationship

Types of Class Invariants

1. Value Constraints (Bounds, Ranges, Allowed Values)

These invariants restrict individual fields to specific domains:

value-constraints
TypeScript
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
class Percentage {
    private value: number;
    
    // INVARIANT: 0 <= value <= 100
    
    constructor(value: number) {
        if (value < 0 || value > 100) {
            throw new Error("Percentage must be between 0 and 100");
        }
        this.value = value;
    }
    
    add(other: Percentage): Percentage {
        // Result capped to maintain invariant
        const sum = Math.min(100, this.value + other.value);
        return new Percentage(sum);
    }
}
 
class EmailAddress {
    private readonly address: string;
    
    // INVARIANT: address is a valid email format (contains @ and domain)
    
    constructor(address: string) {
        if (!this.isValidEmail(address)) {
            throw new Error("Invalid email address format");
        }
        this.address = address;
    }
    
    private isValidEmail(email: string): boolean {
        // Simplified validation
        const pattern = /^[^\s@]+@[^\s@]+\.[^\s@]+$/;
        return pattern.test(email);
    }
}
 
class AgeBracket {
    private minAge: number;
    private maxAge: number;
    
    // INVARIANT: 0 <= minAge <= maxAge <= 150
    
    constructor(minAge: number, maxAge: number) {
        if (minAge < 0 || maxAge > 150 || minAge > maxAge) {
            throw new Error("Invalid age bracket");
        }
        this.minAge = minAge;
        this.maxAge = maxAge;
    }
}

2. Relationship Invariants (Between Fields)

These invariants express that multiple fields must maintain a specific relationship:

relationship-invariants
TypeScript
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
class Rectangle {
    private width: number;
    private height: number;
    private area: number;  // Cached for performance
    
    // INVARIANT: area === width * height (always synchronized)
    
    constructor(width: number, height: number) {
        this.width = width;
        this.height = height;
        this.area = width * height;  // Establish invariant
    }
    
    setWidth(newWidth: number): void {
        this.width = newWidth;
        this.area = this.width * this.height;  // Maintain invariant
    }
    
    setHeight(newHeight: number): void {
        this.height = newHeight;
        this.area = this.width * this.height;  // Maintain invariant
    }
    
    getArea(): number {
        // Relies on invariant - no recalculation needed
        return this.area;
    }
}
 
class SortedList<T> {
    private items: T[];
    private comparator: (a: T, b: T) => number;
    
    // INVARIANT: For all i < j: comparator(items[i], items[j]) <= 0
    // (Items are always in sorted order)
    
    constructor(comparator: (a: T, b: T) => number) {
        this.items = [];
        this.comparator = comparator;
    }
    
    add(item: T): void {
        // Binary search to find insertion point - maintains sort invariant
        let left = 0, right = this.items.length;
        while (left < right) {
            const mid = Math.floor((left + right) / 2);
            if (this.comparator(this.items[mid], item) <= 0) {
                left = mid + 1;
            } else {
                right = mid;
            }
        }
        this.items.splice(left, 0, item);
        // Invariant maintained: items remain sorted
    }
    
    getMin(): T | undefined {
        // Relies on invariant: smallest is always at index 0
        return this.items[0];
    }
    
    getMax(): T | undefined {
        // Relies on invariant: largest is always at last index
        return this.items[this.items.length - 1];
    }
}

3. Structural Invariants (Data Structure Integrity)

These invariants ensure that internal data structures remain valid:

structural-invariants
TypeScript
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
67
68
69
class BinarySearchTree<T> {
    private root: TreeNode<T> | null = null;
    private comparator: (a: T, b: T) => number;
    
    // INVARIANT: For every node N:
    //   - All values in N.left subtree < N.value
    //   - All values in N.right subtree > N.value
    // (Binary Search Tree property)
    
    constructor(comparator: (a: T, b: T) => number) {
        this.comparator = comparator;
    }
    
    insert(value: T): void {
        // Insert maintains BST invariant by choosing correct position
        if (!this.root) {
            this.root = { value, left: null, right: null };
            return;
        }
        
        this.insertRecursive(this.root, value);
        // Invariant maintained: BST property preserved
    }
    
    private insertRecursive(node: TreeNode<T>, value: T): void {
        const comparison = this.comparator(value, node.value);
        
        if (comparison < 0) {
            // Value is less - goes in left subtree (maintains invariant)
            if (node.left) {
                this.insertRecursive(node.left, value);
            } else {
                node.left = { value, left: null, right: null };
            }
        } else {
            // Value is greater/equal - goes in right subtree (maintains invariant)
            if (node.right) {
                this.insertRecursive(node.right, value);
            } else {
                node.right = { value, left: null, right: null };
            }
        }
    }
    
    contains(value: T): boolean {
        // Search relies on BST invariant for O(log n) lookup
        return this.searchRecursive(this.root, value);
    }
    
    private searchRecursive(node: TreeNode<T> | null, value: T): boolean {
        if (!node) return false;
        
        const comparison = this.comparator(value, node.value);
        if (comparison === 0) return true;
        
        // Invariant guarantees we only need to search one subtree
        if (comparison < 0) {
            return this.searchRecursive(node.left, value);
        } else {
            return this.searchRecursive(node.right, value);
        }
    }
}
 
interface TreeNode<T> {
    value: T;
    left: TreeNode<T> | null;
    right: TreeNode<T> | null;
}

4. Count/Size Invariants (Synchronized Counts)

These invariants ensure that size or count fields accurately reflect the actual data:

count-invariants
TypeScript
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
class LinkedList<T> {
    private head: ListNode<T> | null = null;
    private tail: ListNode<T> | null = null;
    private count: number = 0;
    
    // INVARIANT 1: count === actual number of nodes in list
    // INVARIANT 2: if count === 0, then head === null && tail === null
    // INVARIANT 3: if count > 0, then head !== null && tail !== null
    // INVARIANT 4: tail.next === null (tail is always the last node)
    
    add(value: T): void {
        const newNode: ListNode<T> = { value, next: null };
        
        if (this.tail) {
            this.tail.next = newNode;
            this.tail = newNode;
        } else {
            this.head = newNode;
            this.tail = newNode;
        }
        
        this.count++;  // Maintain count invariant
        // All invariants maintained
    }
    
    removeFirst(): T | undefined {
        if (!this.head) return undefined;
        
        const value = this.head.value;
        this.head = this.head.next;
        
        if (!this.head) {
            this.tail = null;  // Maintain invariant 2
        }
        
        this.count--;  // Maintain count invariant
        // All invariants maintained
        
        return value;
    }
    
    size(): number {
        // Relies on count invariant - O(1) instead of traversing
        return this.count;
    }
    
    isEmpty(): boolean {
        // Relies on invariant 2
        return this.count === 0;
    }
}
 
interface ListNode<T> {
    value: T;
    next: ListNode<T> | null;
}

Documenting Invariants

Why Invariants Matter

Invariants might seem like additional overhead—extra checks, extra thinking, extra documentation. But they provide profound benefits that compound as systems grow in size and complexity.

1. Trust and Simplification

When invariants are guaranteed, code can make assumptions without defensive checks:

Without Invariant Guarantees

•Every method must validate inputs defensively
•Callers can't trust returned values
•Business logic drowns in validation code
•Bugs hide in forgotten edge cases
•Code becomes verbose and brittle

With Invariant Guarantees

•Methods assume valid state on entry
•Returned objects are always valid
•Business logic is clear and focused
•Invalid states are prevented by design
•Code is concise and robust

trust-simplification
TypeScript
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
// WITHOUT invariant guarantees - defensive everywhere
class UntrustedRectangle {
    width: number | null = null;
    height: number | null = null;
    
    getArea(): number | null {
        // Must check everything because nothing is guaranteed
        if (this.width === null || this.height === null) {
            return null;
        }
        if (this.width < 0 || this.height < 0) {
            return null;  // Or throw? Caller doesn't know what to expect
        }
        return this.width * this.height;
    }
    
    getPerimeter(): number | null {
        // Same defensive checks repeated everywhere
        if (this.width === null || this.height === null) {
            return null;
        }
        if (this.width < 0 || this.height < 0) {
            return null;
        }
        return 2 * (this.width + this.height);
    }
}
 
// WITH invariant guarantees - clean and focused
class TrustedRectangle {
    // INVARIANT: width > 0, height > 0 (always positive, never null)
    
    private constructor(
        private readonly width: number,
        private readonly height: number
    ) {}
    
    static create(width: number, height: number): TrustedRectangle {
        // Invariant enforced at construction
        if (width <= 0 || height <= 0) {
            throw new Error("Dimensions must be positive");
        }
        return new TrustedRectangle(width, height);
    }
    
    getArea(): number {
        // No checks needed - invariant guarantees valid dimensions
        return this.width * this.height;
    }
    
    getPerimeter(): number {
        // Clean implementation relying on invariant
        return 2 * (this.width + this.height);
    }
    
    scale(factor: number): TrustedRectangle {
        if (factor <= 0) {
            throw new Error("Scale factor must be positive");
        }
        // Result automatically satisfies invariant
        return new TrustedRectangle(this.width * factor, this.height * factor);
    }
}

2. Reasoning About Correctness

Invariants make it possible to reason about program correctness systematically:

During code review: Verify each method maintains invariants
During debugging: If invariant is violated, you know exactly what broke
During testing: Tests can verify invariants hold after each operation
During design: Invariants guide API design and reveal impossible operations

3. Performance Optimization

Invariants enable optimizations that would otherwise be unsafe:

invariant-optimization
TypeScript
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
67
68
69
class SortedArray<T> {
    private items: T[];
    private comparator: (a: T, b: T) => number;
    
    // INVARIANT: items is always sorted according to comparator
    
    constructor(comparator: (a: T, b: T) => number) {
        this.items = [];
        this.comparator = comparator;
    }
    
    /**
     * Because of the invariant, we can use binary search.
     * O(log n) instead of O(n) for unsorted arrays.
     */
    contains(item: T): boolean {
        return this.binarySearch(item) !== -1;
    }
    
    /**
     * Find minimum in O(1) because invariant guarantees smallest is first.
     */
    min(): T | undefined {
        return this.items[0];  // Invariant: first is always smallest
    }
    
    /**
     * Find maximum in O(1) because invariant guarantees largest is last.
     */
    max(): T | undefined {
        return this.items[this.items.length - 1];  // Invariant: last is always largest
    }
    
    /**
     * Merge two sorted arrays in O(n + m) because both maintain sort invariant.
     * Without the invariant, this would require O((n+m) log(n+m)) to sort.
     */
    merge(other: SortedArray<T>): SortedArray<T> {
        const result = new SortedArray<T>(this.comparator);
        let i = 0, j = 0;
        
        // Merge in sorted order - only possible because both are sorted
        while (i < this.items.length && j < other.items.length) {
            if (this.comparator(this.items[i], other.items[j]) <= 0) {
                result.items.push(this.items[i++]);
            } else {
                result.items.push(other.items[j++]);
            }
        }
        
        // Append remaining items (already sorted)
        while (i < this.items.length) result.items.push(this.items[i++]);
        while (j < other.items.length) result.items.push(other.items[j++]);
        
        return result;  // Result also maintains invariant
    }
    
    private binarySearch(item: T): number {
        let left = 0, right = this.items.length - 1;
        while (left <= right) {
            const mid = Math.floor((left + right) / 2);
            const cmp = this.comparator(this.items[mid], item);
            if (cmp === 0) return mid;
            if (cmp < 0) left = mid + 1;
            else right = mid - 1;
        }
        return -1;
    }
}

Breaking Invariants Breaks Optimizations

Invariants as Implicit Contracts

Here's a critical insight: Class invariants create implicit contracts between the class and all code that uses it.

When you define a class with invariants, you're making promises to callers:

"This BankAccount will never have a negative balance"
"This DateRange will always have start ≤ end"
"This SortedList will always be in order"

Callers rely on these promises. They write code that assumes the invariants are true. If you violate the invariants, you break that code—often in subtle, hard-to-detect ways.

implicit-contracts
TypeScript
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
class NonNegativeBalance {
    // INVARIANT: balance >= 0
    private balance: number;
    
    constructor(initialBalance: number) {
        if (initialBalance < 0) {
            throw new Error("Initial balance cannot be negative");
        }
        this.balance = initialBalance;
    }
    
    deposit(amount: number): void {
        if (amount < 0) {
            throw new Error("Cannot deposit negative amount");
        }
        this.balance += amount;
        // Invariant maintained: adding positive to non-negative stays non-negative
    }
    
    withdraw(amount: number): void {
        if (amount < 0) {
            throw new Error("Cannot withdraw negative amount");
        }
        if (amount > this.balance) {
            throw new Error("Insufficient funds");
        }
        this.balance -= amount;
        // Invariant maintained: we checked amount <= balance
    }
    
    getBalance(): number {
        return this.balance;  // Caller can trust this is >= 0
    }
}
 
// Caller code relies on the invariant
function calculateInterest(account: NonNegativeBalance): number {
    const balance = account.getBalance();
    
    // This code TRUSTS that balance >= 0
    // No check needed because the invariant guarantees it
    const rate = 0.05;
    return balance * rate;  // Always non-negative
}
 
function displayBalance(account: NonNegativeBalance): string {
    const balance = account.getBalance();
    
    // Safe to use without checking for negative
    // The invariant guarantees balance >= 0
    return `$${balance.toFixed(2)}`;
}
 
function transferSafely(
    from: NonNegativeBalance, 
    to: NonNegativeBalance, 
    amount: number
): void {
    // This operation is safe because both accounts maintain invariants
    from.withdraw(amount);  // Will throw if insufficient funds
    to.deposit(amount);
    // If we reach here, both accounts still satisfy invariants
}

The Trust Chain:

Invariants create a chain of trust throughout your codebase:

Class guarantees invariant → Methods ensure invariants hold after every operation
Methods rely on invariant → Internal implementation doesn't re-check what's guaranteed
Callers rely on invariant → External code trusts the class promises
Downstream relies on caller → Systems built on top trust their building blocks

Make Invariants Unbreakable

Strategies for Maintaining Invariants
Strategy	How It Works	When to Use
Private fields only	External code cannot directly modify state	Always - default approach
Validate in constructor	Reject invalid initial states	Always - establish invariant at birth
Validate in mutators	Check before every state change	When mutation is necessary
Immutability	Objects can't change, so invariants can't be violated post-construction	When feasible - strongest guarantee
Factory methods	Control object creation to ensure valid states	Complex initialization logic
Defensive copies	Prevent callers from mutating internal state	When returning mutable internals

Identifying Invariants in Existing Code

How do you uncover the invariants in code that doesn't document them? This is a critical skill, especially when working with legacy systems or third-party libraries.

Discovery Techniques:

How to Find Hidden Invariants

•Look at constructors: What validation happens? What must be true for construction to succeed?
•Examine field types: Are they non-nullable? Bounded? What types were chosen?
•Read setter/mutator methods: What throws exceptions? What conditions are checked?
•Study getter methods: What assumptions do they make about internal state?
•Analyze assertions and checks: What conditions does the code verify?
•Check test cases: What properties do tests verify? What edge cases are covered?
•Review documentation: Even informal comments hint at expected properties

discovering-invariants
TypeScript
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
// Given this class, what are its invariants?
class Order {
    private items: OrderItem[];
    private status: OrderStatus;
    private total: number;
    private createdAt: Date;
    private shippedAt: Date | null;
    
    constructor(items: OrderItem[]) {
        if (items.length === 0) {
            throw new Error("Order must have at least one item");
        }
        this.items = [...items];  // Defensive copy
        this.status = OrderStatus.PENDING;
        this.total = this.calculateTotal();
        this.createdAt = new Date();
        this.shippedAt = null;
    }
    
    private calculateTotal(): number {
        return this.items.reduce((sum, item) => sum + item.price * item.quantity, 0);
    }
    
    ship(): void {
        if (this.status !== OrderStatus.PENDING) {
            throw new Error("Can only ship pending orders");
        }
        this.status = OrderStatus.SHIPPED;
        this.shippedAt = new Date();
    }
    
    getItems(): OrderItem[] {
        return [...this.items];  // Defensive copy - can't modify internal
    }
}
 
// DISCOVERED INVARIANTS:
// 1. items.length > 0 (checked in constructor)
// 2. total === sum of item prices * quantities (set at construction)
// 3. createdAt is never null (set at construction)
// 4. If status === SHIPPED, then shippedAt !== null
// 5. If status === PENDING, then shippedAt === null
// 6. items array is immutable after construction (defensive copies)
// 7. status transitions: PENDING -> SHIPPED (never backwards)

Common Invariant Patterns:

Common Invariants in Software Systems
Domain	Common Invariants
Financial	balance >= 0; credits - debits = balance; transaction amounts > 0
Collections	size >= 0; size = actual count; indices < size
Date/Time	start <= end; duration >= 0; times are in valid ranges
Graphs	edge connects valid nodes; no orphan edges; degrees are accurate
Trees	one root; no cycles; parent-child links consistent
Caches	capacity >= 0; size <= capacity; entries match stored keys
User Data	id is unique; email format is valid; required fields are non-null

Document What You Discover

Summary: The Foundation of Object Integrity

We've explored class invariants in depth. Let's consolidate the key insights:

Key Takeaways

•Invariants are always-true conditions — Properties that must hold from object creation through its entire lifetime
•Invariants come in many forms — Value constraints, field relationships, structural properties, synchronized counts
•Every method must maintain invariants — Methods can temporarily break invariants during execution, but must restore them before returning
•Invariants enable trust and optimization — Code can assume invariant properties without defensive checks
•Invariants create implicit contracts — Callers rely on invariant properties; violating them breaks downstream code
•Document your invariants explicitly — Future maintainers must understand what properties the class guarantees

What's Next:

You'll learn exactly how subclasses can accidentally break parent invariants, and how to design inheritance hierarchies that maintain invariant integrity across the entire type hierarchy.

Page Complete

1 / 4