When should you use a computed property versus a method or a stored property?

Answer: Use a computed property for a cheap, side-effect-free value derived from existing state; use a method for expensive work, side effects, or anything that needs arguments.

Choose	When
Stored property	The value is its own state (not derivable), or caching an expensive result is worth the sync risk
Computed property	Derived from other state, O(1)-ish, no side effects, reads like an attribute (isEmpty, fullName)
Method	Expensive (O(n) or I/O), has side effects, may fail, or takes parameters (fetchUser(), distance(to:))
lazy var	Expensive, needed at most once, and the inputs never change afterwards

Code Example:

struct Order {
    var items: [Item]

    var isEmpty: Bool { items.isEmpty }                 // ✅ cheap, derived
    var itemCount: Int { items.count }                  // ✅ cheap, derived

    func totalPrice(applying discount: Discount) -> Decimal { // ✅ method — takes input
        items.reduce(0) { $0 + $1.price } * discount.multiplier
    }
}

Key Points:

• Readers assume property access is fast — a computed property that hits disk or the network is a surprise.
• Caching a derived value in a stored property creates two sources of truth that can drift apart; prefer computed unless profiling says otherwise.