Coupling measures how much one module depends on another; cohesion measures how strongly the parts inside a module belong together. The goal: low coupling, high cohesion.
HIGH cohesion → a class does ONE well-defined job, parts relate strongly ✅
LOW coupling → modules interact through small, stable interfaces ✅
# LOW cohesion: unrelated responsibilities jammed together
class Utils:
def parse_csv(self, s): ...
def send_email(self, to): ...
def hash_password(self, p): ... # nothing in common
# HIGH cohesion: each class has a single, focused purpose
class CsvParser: ...
class Mailer: ...
class PasswordHasher: ...
# TIGHT: OrderService is welded to a concrete class
class OrderService:
def __init__(self):
self.db = MySqlDatabase() # can't swap or test in isolation
# LOOSE: depends on an abstraction, injected in
class OrderService:
def __init__(self, db: Database): # any Database works
self.db = db
Loose coupling lets you replace MySqlDatabase with an in-memory fake for tests, with no edit to OrderService.
Tight coupling is what makes a change "ripple" — touch one class and ten others break; loose coupling contains changes to their source.
High cohesion makes code findable and testable: when a class does one thing, you know where logic lives and what a test should cover.
These two metrics underlie nearly every other design principle.
A library of IT interview questions with detailed answers — from Junior to Senior.
Donate