Stage 1 of 10 · 8 lessons · ~3 weeks

Object-oriented programming properly

Go beyond the textbook definitions: use classes to protect rules, interfaces to hide change, polymorphism to remove type checks and composition instead of deep inheritance, in your language's modern style.

Start: Classes, objects and responsibilities

Lessons· one lesson for each topic, in this order

Do this

Rewrite one old project class so that all fields are private and every rule is enforced in a method.
Replace one if or switch on a type field with an interface and two classes.
Refactor one inheritance you wrote for code reuse into composition.
Write a small immutable value type, such as Money or Email, with its own validation.
Write unit tests that prove your invariants cannot be broken from outside.

Start with· follow one of these

CoursePaid
Grokking the Object Oriented Design Interview
Design Gurus · Intermediate · Subscription
Its opening chapters teach the four OOP ideas applied inside real designs, with UML and running code
CourseFree
MIT 6.102: Software Construction
MIT · Intermediate
Free MIT readings: Abstract Data Types, Rep Invariants and Interfaces and Subtyping are this stage at university depth

Practise

Exercism
Free exercises in your language with optional human mentoring on design, not just on passing tests

Keep handy

The Python Tutorial
Read chapter 9, Classes, then the dataclasses and typing pages in the standard library docs

Also worth your time

Beginner's Guide to Object-Oriented Programming (OOP) (opens in a new tab)Read
The four pillars with one running example, a quick way to check your definitions before going deeper
A Philosophy of Software DesignGo deeperPaid
Deep modules and information hiding: the clearest book on why encapsulation matters

Before moving on: A small domain model you wrote from a one-paragraph brief, such as a library or a clinic, with private state, at least one interface with two implementations, no type switches and composition wherever inheritance was tempting, plus unit tests for its rules.

Tauseef's advice

Python lets you skip classes entirely, and often you should: a function or a dataclass is frequently the better design. Use classes when there is state with rules to protect, and use Protocol for interfaces so a type checker can check your design.