Separation of concerns divides a system (via divide-and-conquer) into units with minimal overlap; encapsulation places data + its operations into a unit and exposes only a controlled interface.

Divide and conquer: split the problem, solve smaller independent problems, integrate the solutions.

Abstraction: hide and generalise details to reduce complexity (level differs: use-case model vs code).

Encapsulation / modularity / information hiding: put relevant data into a unit and expose it only via a few public methods.

A 'unit' can be a method, class, or package depending on the abstraction level.

Coupling is the degree of interaction between units; cohesion is the degree to which a unit fulfils one purpose. Design goal: minimise coupling, maximise cohesion.

Information hiding: reveal as little as needed; prevent data manipulation; ensure unit integrity; expose a single controlled interface — done via visibility (public/package/protected/private) and interfaces.

The interface to a class is its public methods.

DRY (Don't Repeat Yourself): solve each problem in one place; avoid copy-paste clones; unify similar code via parameters/auxiliary methods.

Good separation: one unit = one well-defined problem; small interfaces (few parameters, pass only needed data, combine into reference types); small/clear methods.