DEV Community

Cover image for The Art of Object-Oriented Programming
The Silicon Architect profile image Syed Ahmer Shah
Syed Ahmer Shah for The Silicon Architect

Posted on Originally published at Medium AI-assisted

The Art of Object-Oriented Programming

#ai #webdev #programming #productivity

Most developers learn OOP backwards — memorize four buzzwords, pass the exam, then go write procedural code inside classes. That’s not object-oriented programming. That’s just spaghetti with labels on it.

OOP is a design philosophy before it’s a syntax. It’s a way of breaking a complex system into smaller, self-contained pieces — pieces that each know their job, protect their own state, and can talk to each other without the whole thing falling apart. The syntax is easy. The thinking is what takes years.

Classes & Objects

A class is a blueprint. An object is the thing built from it.

Think of an architect’s floor plan for a house. The blueprint describes two bedrooms, a kitchen, a front door. But you can’t live in the blueprint. When a builder follows it and constructs an actual house — that’s the object. Build fifty houses from the same blueprint and each one is its own thing: different paint, different furniture, different family inside.

Same class. Two separate objects. This is the starting point.

State, Behavior & Identity

Every object carries three things: what it knows, what it can do, and what makes it uniquely itself.

State is the data — a bank account’s balance, a user’s email address. Behavior is the methods — a bank account can deposit, withdraw, generate a statement. Identity is what separates one object from another even if their data is identical. Two bank accounts with the same zero balance are still two separate accounts.

A constructor is where this all begins. It’s the object’s moment of birth — where you set up valid, usable initial state before anyone else touches it.

Good constructors do one thing: produce a valid object and get out. No database calls, no network requests, no side effects. Just state.

Methods are where responsibility lives — but the most important question about any method isn’t “does it work?” It’s “does it belong here?” A User class should not have a method called sendInvoice(). That's not the user's job. When a method doesn't belong, it's a signal that your boundaries are wrong.

The Four Pillars — What They Actually Mean

Encapsulation

Encapsulation is not “make things private.” That’s the mechanism. The principle is protecting your state from being changed in ways you didn’t intend.

Think of a vending machine. You can see the snacks through the glass, put money in, and press a button. But you cannot reach inside and grab whatever you want. The machine controls its own internals. You interact through a defined interface.

The stock field is private. No external code can set stock = 999. All changes go through purchase(), which enforces the rules. That's the point — not hiding data for its own sake, but controlling how it changes.

Abstraction

Abstraction hides complexity behind a simple interface.

When you drive a car, you don’t need to understand internal combustion to press the accelerator. The car abstracts all of that away. You get a pedal. You press it. The machinery underneath is none of your business.

The calling code just calls .charge(). It doesn't care how. Tomorrow you add a PayPal gateway. The calling code doesn't change. That's abstraction doing its job.

Inheritance

Inheritance captures the “is-a” relationship. A Dog is an Animal. A SavingsAccount is a BankAccount. You inherit the shared structure, then specialize.

Here’s where most people go wrong: they use inheritance for code reuse rather than for actual “is-a” relationships. If you’re inheriting from a class just to borrow two methods, you’ve built a dependency that’ll betray you when the parent class needs to change. Inheritance should model a genuine relationship — not be a shortcut.

Polymorphism

Polymorphism means the same interface, different behavior. One method name, many implementations depending on which object you’re actually calling it on.

printArea doesn't know or care whether it's working with a circle or a rectangle. It calls .area() and trusts the object to do the right thing. Add a Triangle tomorrow — printArea doesn't need to touch it. This is what makes systems extensible without becoming fragile.

Composition vs. Inheritance — The Choice Nobody Teaches Properly

Real OOP design lives in this question: should this object be the other thing, or should it have it?

A Car doesn't inherit from Engine. That would be absurd. A Car has an engine. It uses an engine. That's composition.

If tomorrow you need an ElectricEngine, you swap it in. The Car class doesn't change at all. Inheritance would have made that painful. Composition keeps it trivial.

The general rule: prefer composition. Use inheritance only when you have a genuine “is-a” relationship and you want to express that relationship explicitly in the design.

Association, Aggregation & Composition

These three describe how tightly objects are bound to each other — and it matters more than most tutorials acknowledge.

Association is the loosest connection. A Doctor treats Patients. They know about each other. Neither owns the other. They can both exist independently.

Aggregation is a whole-part relationship where the parts survive without the whole. A Team has Players. Dissolve the team, the players still exist somewhere.

Composition is the tightest bond — the part cannot exist without the whole. A House has Rooms. Destroy the house, the rooms aren't standalone things anymore.

Getting this right isn’t academic. It determines how you structure your database schemas, how you handle object lifecycles, and what breaks when something gets deleted.

Coupling & Cohesion — The Two Dials

Every system has two dials you’re constantly tuning.

Coupling measures how much one class depends on another. High coupling means a change in one class ripples into others, sometimes into five others you didn’t expect. Low coupling means classes can change independently. You want low coupling.

Cohesion measures how focused a class is on a single purpose. A class that handles user registration, sends confirmation emails, generates PDFs, and logs activity — that’s low cohesion. It’s doing four different jobs. When you split that into four focused classes, each becomes easier to test, easier to change, and easier to understand.

High cohesion and low coupling together produce modular code. A UserRegistrationService that only handles registration and receives its email sender and logger through injection rather than creating them — that's what well-designed OOP looks like in practice.

The question to ask every time you write a class: if I had to change this one thing, how many other places would break?

Common OOP Design Mistakes

A few patterns that feel right until they don’t:

The God Object. One massive class that knows everything and controls everything. Usually named Manager, Handler, or Utilities. This class grows without limit because "it's already there." Split it ruthlessly.

Anemic domain models. Objects that are just bags of data — nothing but getters and setters — with all the actual logic sitting in a separate service. This is procedural code wearing OOP clothes. Behavior belongs with the data it operates on.

Deep inheritance chains. Five levels deep, the sixth developer can’t understand what they’re actually working with. The deeper the chain, the harder it breaks. Flatten it with composition.

Abstracting too early. Building interfaces and abstract base classes before you have two real concrete variations. Write the concrete thing first. Abstract only when you see the pattern emerge from actual code — not from prediction.

A Small Real-World Example

Consider an order system. A bad design puts everything into Order: it calculates totals, applies discounts, sends emails, and updates inventory. Every time any one of those things changes, you're editing this one class.

A better design:

Three classes. Three clear jobs. Order knows its own data. DiscountService knows how pricing works. OrderNotifier knows how to reach the customer. None of them know too much about the others. Swap email notifications for SMS tomorrow — Order doesn't care, DiscountService doesn't care. Only OrderNotifier changes.

This is low coupling and high cohesion working together. It’s not clever. It’s disciplined.

Conclusion: OOP as a Design Mindset

OOP isn’t about memorizing four words. It’s not a style guide. It’s a discipline for managing complexity — for breaking a system that’s too big to hold in your head into pieces that each have a clear owner and a clear job.

The programmers who use it well think before they type. They ask: where does this knowledge live? Who is responsible for this behavior? What happens to the rest of the system if this changes? They treat their objects the way an architect treats walls — with respect for what they hold up.

Learn the concepts once. Then spend the rest of your time learning how they work together. The syntax is a day’s work. The design sense takes years. That’s where the actual art is.

References

  1. Gamma, Helm, Johnson, Vlissides — Design Patterns: Elements of Reusable Object-Oriented Software

  2. Martin Fowler — Refactoring: Improving the Design of Existing Code

  3. Robert C. Martin — Clean Code: A Handbook of Agile Software Craftsmanship

  4. MDN Web Docs — Object-Oriented Programming in JavaScript

  5. Mattias Petter Johansson — Composition over Inheritance (FunFunFunction, YouTube)

Find me across the web:

Top comments (0)