The Object Class, equals/hashCode, and SOLID

I. Notebook Summary (copy by hand)

Terse, high-density. Keyword → rule. Everything here is derivable from Section II.

1 · The Object class (the "godfather")

2 · toString()

3 · equals()

instanceof vs getClass() in equals() [+] interview favourite

4 · hashCode()

5 · String hashCode & the magic 31

6 · SOLID (Uncle Bob / Robert C. Martin)

II. Detailed Study Notes

Built from scratch, beginner-friendly. [+] marks additions/corrections beyond the core material.

1. The Object class — the root of everything

Every Java class silently extends java.lang.Object, so a handful of methods are available on every object you ever touch.

Java has a single rooted inheritance hierarchy. When you write class Student {}, the compiler treats it as class Student extends Object {}. If your class already extends something else, that parent (or its parent, eventually) extends Object. So Object is the ancestor of all classes — directly or indirectly. It lives in the java.lang package, which is auto-imported.

Because of this, every object inherits these methods for free:

→ Code: Object.java

2. toString() — making objects printable

Printing an object reference secretly calls its toString(); the default output is cryptic, so you override it.

When you run System.out.println(student), Java doesn't print the object's memory; it implicitly calls student.toString() and prints that string. The default implementation in Object is:

public String toString() {
    return getClass().getName() + "@" + Integer.toHexString(hashCode());
}

So an un-overridden object prints like com.demo.Student@1b6d3586. Reading left to right: getClass().getName() is the fully-qualified class name, "@" is a literal separator, and Integer.toHexString(hashCode()) is the object's hash code rendered in hexadecimal. [+] That hex string is not a memory address — a common misconception. It's just the hashCode in base-16; on the default implementation it's derived from the object's identity, but the JVM does not expose raw pointers.

Override toString() to return something meaningful:

@Override
public String toString() {
    return "Student{firstName=" + firstName + ", lastName=" + lastName + ", email=" + email + "}";
}

Why override it: meaningful one-line printing via println; invaluable for logging and debugging; represents all fields (including numeric ones) as a single string. Many built-in classes already override it — the wrapper classes (Integer, Long, Double, Boolean, Character, etc.), the collection classes, and String/StringBuilder/StringBuffer — which is why printing a list shows [1, 2, 3] rather than gibberish.

→ Code: ToStringDemo.java · Lab: Exercise 1

3. equals() — value equality vs identity

By default equals() checks whether two references point to the same object; override it to compare by the fields you care about.

Create two Student objects with identical first name, last name, and email, and a.equals(b) still returns false. Reason: the inherited Object.equals() is literally this == obj — a reference (memory-location) comparison. Two separately-constructed objects live at different locations, so they're "not equal" even when their data matches.

To make equality mean "same student id", override it:

@Override
public boolean equals(Object obj) {
    if (obj == null) return false;              // nothing equals null
    if (!(obj instanceof Student)) return false; // wrong type -> not equal
    if (obj == this) return true;                // same reference -> fast yes
    return this.getId() == ((Student) obj).getId(); // compare by id
}

Read it as a funnel: reject null, reject wrong types, accept an identical reference instantly, then fall back to the real field comparison. The instanceof check also doubles as a type guard so the cast on the last line is safe. [+] Note instanceof already returns false for null, so the explicit null check above is technically redundant when you use instanceof — it's kept here for clarity.

→ Code: EqualsDemo.java · Lab: Exercise 2

4. instanceof vs getClass() inside equals() [+]

The type check at the top of equals() can use instanceof or getClass(), and the choice quietly changes how subclasses behave — a favourite interview trap.

This wasn't spelled out in my core material, but it's the natural follow-up question, so it's worth nailing down with verified reasoning:

Joshua Bloch (Effective Java) favours instanceof, because getClass() means an object can only ever equal another of the exact same runtime class — so a harmless subclass that merely adds a method is suddenly "not equal" to its parent, which is surprising and brushes against the Liskov principle. The catch with instanceof: if a subclass overrides equals and folds in extra fields, you can break symmetry (parent says "equal", child says "not equal"). The standard fix is to make equals (and hashCode) final, or to only use instanceof when subclasses don't change the notion of equality.

→ Code: InstanceofVsGetClass.java

5. hashCode() — the address into a hash table

Hash-based collections find objects by their hashCode() first; if you override equals() but not hashCode(), a HashSet will store visible duplicates.

Store two "equal" students (you've already overridden equals) in a HashSet and call set.contains(other) — it can still return false. Why, if a HashSet stores unique objects? Because a HashSet doesn't scan every element. It keeps elements in buckets, and each bucket is chosen from the object's hash code. The lookup goes: compute hashCode() → jump to that bucket → only there compare with equals().

If two equal students return different hash codes (the default behaviour, since default hashCode is per-instance), they land in different buckets and never get compared — so the set never notices they're duplicates. Fix it by making equal objects produce equal hash codes:

@Override
public int hashCode() {
    return id;   // equal ids -> equal hashCode -> same bucket
}

→ Code: HashCodeDemo.java · Lab: Exercise 3

6. The equals/hashCode contract

Equality and hashing follow a small set of rules; breaking them is how collections silently corrupt.

equals() contract — for non-null references:

• Reflexive: x.equals(x) is true.
• Symmetric: x.equals(y) iff y.equals(x).
• Transitive: if x=y and y=z then x=z.
• Consistent: repeated calls give the same result if no fields change.
• x.equals(null) is always false.

The hashCode bridge — the one rule everyone forgets:

So can two unequal objects share a hash code? Yes — that's a collision, and it's allowed. But two equal objects must never disagree on their hash code. [+] Mnemonic: "equals implies same hash, but same hash does not imply equals."

7. Collisions and how HashSet actually stores things

A bad hashCode() doesn't break correctness — it wrecks performance by piling everything into a few buckets.

Replace return id; with return id % 5; and loading the same data into a HashSet takes noticeably longer. The reason: id % 5 yields only five distinct hash codes, so thousands of students cram into five buckets. Within a bucket, finding/inserting means comparing against everything already there — so you've traded near-instant lookups for slow scans. The lesson: override hashCode() wisely to spread values and avoid collisions.

[+] How a HashSet works under the hood (verified): a HashSet is backed by a HashMap. Insertion:

1. Compute hashCode() of the element.
2. Map it to a bucket index — conceptually hash % capacity, but actually hash & (capacity-1) since capacity is a power of two. [+]
3. Empty bucket → store it. Occupied bucket (collision) → compare via equals(); if no match, chain it in.

[+] There's also resizing: when the set fills past its load factor (default 0.75), capacity doubles and everything is rehashed — which is partly why repeatedly adding to a poorly-distributed set is slow.

→ Code: CollisionDemo.java · Lab: Exercise 4

8. Why String's hashCode uses 31

String.hashCode() is a polynomial in 31; the constant is an odd prime chosen mostly because it's cheap to compute, not because it's mathematically optimal.

The formula (from the Java spec):

s[0]*31^(n-1) + s[1]*31^(n-2) + ... + s[n-1]
// computed iteratively as:  h = 31*h + s[i]

Why 31 specifically (verified against Bloch's Effective Java and the original OpenJDK reasoning):

• Odd: if the multiplier were even and the multiplication overflowed, low bits would be lost (multiplying by an even number is like left-shifting, which drops information).
• Prime: traditional; tends to distribute reasonably. [+] The benefit of "prime" here is fuzzier than people claim — Bloch himself noted 33 (composite) performed almost identically.
• Fast: the JVM can rewrite 31 * i as (i << 5) - i — a shift plus a subtract, cheaper than a multiply on older/RISC hardware.

→ Code: StringHash31.java · Lab: Exercise 5

9. SOLID — five principles of good OO design

SOLID is five design principles from Robert C. Martin ("Uncle Bob") that keep object-oriented code clean, loosely coupled, and easy to change.

Why SOLID: to write clean and maintainable code, to preserve good design over time, to keep modules loosely coupled, and to improve reusability and testability.

10. Single Responsibility Principle (SRP) [+] expanded

A class should do one job; if you can describe it with "and," it probably has too many responsibilities.

My core material named SRP but didn't develop it, so here's the full picture (verified, standard definition). "One reason to change" means each class answers to a single actor/concern. A class that parses input, applies business rules, and writes to a database has three reasons to change — a format change, a rule change, or a storage change — and any of them can break the others.

Analogy: a Swiss-army knife is convenient until one tool jams and you can't trust the rest. Separate tools fail independently. Real-world example: split a Report class into ReportData (holds data), ReportFormatter (renders it), and ReportSaver (persists it). Now changing the export format never risks the persistence logic.

→ Code: SRP_Before / SRP_After · Lab: Exercise 6

11. Open–Closed Principle (OCP) [+] expanded

You should be able to add new behaviour by adding new code, not by editing code that already works.

Also named-but-not-developed in my core material, so here's the verified standard treatment. "Open for extension, closed for modification" means: when a new requirement arrives, you extend the system (new class, new implementation of an interface) rather than reopening and rewriting tested classes. The usual vehicle is polymorphism — program to an interface, then add new implementations.

Classic smell: a growing if/else or switch on a type code. Each new type forces an edit to the same method. Fix: define an interface (e.g. Shape.area()), implement it per type (Circle, Square), and let callers depend on Shape. Adding a Triangle means writing one new class — the area-calculating code never changes.

→ Code: OCP_Before / OCP_After

12. L, I, D — quick preview [+]

My core material focused on S and O; for completeness, the remaining three (verified standard definitions):

• Liskov Substitution (L): any subclass must work correctly wherever its parent is used. Classic violation: a Square extending Rectangle where setting width also changes height — code that trusts Rectangle's behaviour breaks.
• Interface Segregation (I): don't force a class to implement methods it doesn't need. Split a fat Machine interface (print, scan, fax) so a simple printer isn't forced to stub out fax().
• Dependency Inversion (D): high-level modules and low-level modules should both depend on abstractions. Inject a MessageSender interface instead of hard-wiring EmailSender, so you can swap in SMS or a test double without touching the high-level logic.

III. Code Samples (runnable)

Standard library only, Java 17+ (any modern JDK works). Each: filename, run command, comments, expected output.

// Reusable model. Other demos build on this file.
public class Student {
    private int id;
    private String firstName, lastName, email;

    public Student(int id, String firstName, String lastName, String email) {
        this.id = id; this.firstName = firstName;
        this.lastName = lastName; this.email = email;
    }
    public int getId() { return id; }
    public String getEmail() { return email; }
}

public class ToStringDemo {
    static class Plain { int id = 7; }            // no toString override

    static class Nice {
        int id = 7;
        @Override public String toString() {      // meaningful override
            return "Nice{id=" + id + "}";
        }
    }
    public static void main(String[] a) {
        System.out.println(new Plain()); // default: ClassName@hexHashCode
        System.out.println(new Nice());  // calls our override implicitly
    }
}

public class EqualsDemo {
    static class Student {
        int id;
        Student(int id) { this.id = id; }
        @Override public boolean equals(Object obj) {
            if (obj == null) return false;
            if (!(obj instanceof Student)) return false;
            if (obj == this) return true;
            return this.id == ((Student) obj).id; // value equality by id
        }
    }
    public static void main(String[] a) {
        Student s1 = new Student(1), s2 = new Student(1), s3 = new Student(2);
        System.out.println(s1 == s2);       // false: different references
        System.out.println(s1.equals(s2));  // true : same id
        System.out.println(s1.equals(s3));  // false: different id
    }
}

public class InstanceofVsGetClass {
    static class A {}
    static class B extends A {}
    public static void main(String[] x) {
        Object oA = new A(), oB = new B();
        // instanceof: a subclass IS-A superclass
        System.out.println(oB instanceof A);              // true
        // getClass: exact runtime type only
        System.out.println(oB.getClass() == A.class);     // false
        System.out.println(oB.getClass() == B.class);     // true
    }
}

import java.util.HashSet;
public class HashCodeDemo {
    static class Student {
        int id;
        Student(int id) { this.id = id; }
        @Override public boolean equals(Object o) {
            return o instanceof Student && ((Student)o).id == id;
        }
        // Comment OUT the next method to see the bug:
        @Override public int hashCode() { return id; }
    }
    public static void main(String[] a) {
        HashSet<Student> set = new HashSet<>();
        set.add(new Student(1));
        set.add(new Student(1));            // "equal" duplicate
        System.out.println(set.size());     // 1 with hashCode, 2 without
        System.out.println(set.contains(new Student(1))); // true / false
    }
}

import java.util.HashSet;
public class CollisionDemo {
    static class Key {
        int id; boolean bad;
        Key(int id, boolean bad) { this.id = id; this.bad = bad; }
        @Override public boolean equals(Object o) {
            return o instanceof Key && ((Key)o).id == id;
        }
        @Override public int hashCode() {
            return bad ? id % 5 : id;  // bad: only 5 buckets
        }
    }
    static long load(boolean bad, int n) {
        long t = System.nanoTime();
        HashSet<Key> s = new HashSet<>();
        for (int i = 0; i < n; i++) s.add(new Key(i, bad));
        return (System.nanoTime() - t) / 1_000_000; // ms
    }
    public static void main(String[] a) {
        int n = 200_000;
        System.out.println("good hashCode: " + load(false, n) + " ms");
        System.out.println("bad  hashCode: " + load(true,  n) + " ms");
    }
}

public class StringHash31 {
    // Reproduce String.hashCode() by hand
    static int myHash(String s) {
        int h = 0;
        for (int i = 0; i < s.length(); i++) h = 31 * h + s.charAt(i);   // h = (h<<5) - h + c
        return h;
    }
    public static void main(String[] a) {
        System.out.println(myHash("Java"));        // matches built-in
        System.out.println("Java".hashCode());
        // 31*x optimization identity:
        int x = 12345;
        System.out.println((31 * x) == ((x << 5) - x)); // true
    }
}

public class SRP_Demo {
    // BEFORE: one class with 3 reasons to change (data + format + save)
    static class ReportBad {
        String data = "sales=100";
        String toHtml() { return "<p>" + data + "</p>"; }
        void save()     { System.out.println("saving: " + data); }
    }
    // AFTER: one responsibility each
    static class ReportData   { String data = "sales=100"; }
    static class HtmlFormatter { String format(ReportData r){ return "<p>"+r.data+"</p>"; } }
    static class FileSaver     { void save(ReportData r){ System.out.println("saving: "+r.data);} }

    public static void main(String[] a) {
        ReportData d = new ReportData();
        System.out.println(new HtmlFormatter().format(d)); // change format here only
        new FileSaver().save(d);                           // change persistence here only
    }
}

public class OCP_Demo {
    // BEFORE: adding a shape forces editing this method (closed? no).
    // double area(Object s){ if (s instanceof Circle)... else if ... }

    // AFTER: open for extension via an interface, closed for modification.
    interface Shape { double area(); }
    static class Circle implements Shape {
        double r; Circle(double r){this.r=r;}
        public double area(){ return Math.PI*r*r; }
    }
    static class Square implements Shape {
        double s; Square(double s){this.s=s;}
        public double area(){ return s*s; }
    }
    // Adding Triangle = new class, ZERO edits to this method:
    static double total(Shape[] shapes){
        double t=0; for (Shape s: shapes) t += s.area(); return t;
    }
    public static void main(String[] a){
        System.out.printf("%.2f%n", total(new Shape[]{ new Circle(1), new Square(2) }));
    }
}

IV. Hands-On Lab (practice)

Warm-up → Core → Challenge. Predict before you run; reveal the solution only after.

# 1. Check the JDK (need 11+, 17+ recommended)
java -version
javac -version

# 2. Make a working folder
mkdir solid-lab && cd solid-lab

# 3. Compile & run any file (single-file source mode, JDK 11+):
java Exercise1.java          # runs directly, no separate javac needed
# or the classic two-step:
javac Exercise1.java && java Exercise1

Exercise 1 — Print yourself Warm-up

Goal: see the default toString(), then override it.

1. Create Exercise1.java with a class Book holding title and pages.
2. In main, print a Book instance with no override.
3. Add an @Override toString() and print again.

public class Exercise1 {
    static class Book {
        String title; int pages;
        Book(String t, int p){ title=t; pages=p; }
        @Override public String toString(){    // comment out to see default
            return "Book{title=" + title + ", pages=" + pages + "}";
        }
    }
    public static void main(String[] a){
        System.out.println(new Book("SOLID", 320));
    }
}

Exercise 2 — Equality by field Warm-up

Goal: make two distinct objects compare equal by a chosen field.

1. Class User with an email field.
2. Create two users with the same email; print == and .equals() before overriding.
3. Override equals() to compare emails (case-insensitive). Re-test.

public class Exercise2 {
    static class User {
        String email; User(String e){ email=e; }
        @Override public boolean equals(Object o){
            if (!(o instanceof User)) return false;       // also handles null
            return email.equalsIgnoreCase(((User)o).email);
        }
    }
    public static void main(String[] a){
        User u1 = new User("A@x.com"), u2 = new User("a@x.com");
        System.out.println(u1 == u2);       // false (always: 2 objects)
        System.out.println(u1.equals(u2));  // true  (case-insensitive match)
    }
}

Exercise 3 — The HashSet duplicate trap Core

Goal: reproduce the "equal objects still duplicate" bug, then fix it.

1. Reuse User from Exercise 2 (with equals overridden, no hashCode).
2. Add two equal users to a HashSet; print size().
3. Add @Override hashCode() returning email.toLowerCase().hashCode(). Re-test.

import java.util.HashSet;
public class Exercise3 {
    static class User {
        String email; User(String e){ email=e; }
        @Override public boolean equals(Object o){
            return o instanceof User && email.equalsIgnoreCase(((User)o).email);
        }
        @Override public int hashCode(){            // remove to see the bug
            return email.toLowerCase().hashCode();  // equal -> same bucket
        }
    }
    public static void main(String[] a){
        HashSet<User> s = new HashSet<>();
        s.add(new User("a@x.com")); s.add(new User("A@x.com"));
        System.out.println(s.size());   // 1 with hashCode, 2 without
    }
}

Exercise 4 — Measure a bad hashCode Core

Goal: feel the performance cost of collisions.

1. Make a Key with an int id and a switchable hashCode (id vs id % 4).
2. Time loading 200,000 keys into a HashSet for each version (System.nanoTime()).
3. Print both timings.

import java.util.HashSet;
public class Exercise4 {
    static class Key {
        int id; boolean bad; Key(int id, boolean bad){ this.id=id; this.bad=bad; }
        @Override public boolean equals(Object o){ return o instanceof Key && ((Key)o).id==id; }
        @Override public int hashCode(){ return bad ? id % 4 : id; }
    }
    static long run(boolean bad){
        long t=System.nanoTime(); HashSet<Key> s=new HashSet<>();
        for (int i=0;i<200_000;i++) s.add(new Key(i,bad));
        return (System.nanoTime()-t)/1_000_000;
    }
    public static void main(String[] a){
        System.out.println("good: "+run(false)+" ms");
        System.out.println("bad : "+run(true) +" ms");
    }
}

Exercise 5 — Reimplement String.hashCode Core

Goal: match the JDK's String hash and prove the 31 shortcut.

1. Write myHash(String) using h = 31*h + ch.
2. Compare it to "Hello".hashCode() for several strings.
3. Assert 31*x == (x<<5)-x for some x.

public class Exercise5 {
    static int myHash(String s){
        int h=0; for (int i=0;i<s.length();i++) h = 31*h + s.charAt(i); return h;
    }
    public static void main(String[] a){
        for (String s : new String[]{"Hello","SOLID","a"}) System.out.println(s+": mine="+myHash(s)+" jdk="+s.hashCode());
        int x=987; System.out.println("shift trick: " + (31*x == (x<<5)-x));
    }
}

Exercise 6 — Refactor to SRP + OCP Challenge

Goal: take a fat class with a type-switch and refactor it to satisfy both principles.

1. Start from a PaymentProcessor with process(String type, double amt) that does if "card"... else if "cash"... and also logs.
2. SRP: pull logging into its own class/interface.
3. OCP: replace the type-switch with a PaymentMethod interface + one class per method.
4. Add a WalletPayment without editing the dispatch code — prove OCP.

public class Exercise6 {
    // SRP: logging is its own responsibility
    interface Logger { void log(String m); }
    static class ConsoleLogger implements Logger {
        public void log(String m){ System.out.println("[log] " + m); }
    }
    // OCP: each payment type is a separate implementation
    interface PaymentMethod { void pay(double amt); }
    static class CardPayment implements PaymentMethod {
        public void pay(double a){ System.out.println("card -> " + a); }
    }
    static class CashPayment implements PaymentMethod {
        public void pay(double a){ System.out.println("cash -> " + a); }
    }
    static class WalletPayment implements PaymentMethod {   // ADDED, no edits elsewhere
        public void pay(double a){ System.out.println("wallet -> " + a); }
    }
    // Dispatch depends on the abstraction; never changes when types are added
    static class Processor {
        Logger logger;
        Processor(Logger l){ logger = l; }
        void process(PaymentMethod m, double amt){
            m.pay(amt); logger.log("processed " + amt);
        }
    }
    public static void main(String[] a){
        Processor p = new Processor(new ConsoleLogger());
        p.process(new CardPayment(),   100);
        p.process(new CashPayment(),    50);
        p.process(new WalletPayment(),  25);  // works with zero dispatch edits
    }
}

★ Key Takeaways

• Object is the universal root. Every class inherits toString, equals, hashCode, getClass, and more.
• toString() is auto-called by println; override it for readable logging/debugging. Default is ClassName@hexHashCode (hex = hashCode, [+] not a memory address).
• equals() defaults to reference equality; override for value equality and respect the contract (reflexive, symmetric, transitive, consistent, false-on-null).
• instanceof vs getClass() [+] is a real trade-off: instanceof respects LSP but can break symmetry; getClass() is symmetric but rejects subclasses.
• hashCode() & buckets: override it whenever you override equals. Equal objects must share a hashCode or hash collections fail.
• Collisions cost speed, not correctness. Spread hash codes; [+] Java 8+ treeifies fat buckets (O(n)→O(log n)) once a bin passes 8 entries and the table has at least 64 buckets — under 64 it resizes instead — and resizes past load factor 0.75. But treeify only pays off when the keys are Comparable: a tree bin orders by hash, and if every hash collides it needs compareTo to break the tie. Exercise 4 measures exactly this — a non-Comparable key stays quadratic.
• The 31 multiplier in String.hashCode is odd + prime + cheap ((i<<5)-i); [+] traditional, not proven optimal.
• SOLID = Single Responsibility, Open–Closed, Liskov Substitution, Interface Segregation, Dependency Inversion — for clean, loosely-coupled, testable, maintainable code. [+] Guidelines, not laws; don't over-apply.

[+] = my own additions/corrections, all fact-checked. Core spine: Object methods (toString/equals/hashCode) + intro to SOLID with SRP and Open–Closed.