What is structured concurrency and why does it matter?

Deep dive

Structured concurrency means: every coroutine has a parent scope/job, so its lifetime is bounded and observable.

What you get:

• Cancellation propagates (cancel parent => cancel children).
• Failures propagate predictably (one child failing can cancel siblings, unless you use `supervisorScope`).
• No “orphan” background work leaking beyond lifecycle.

Practical guidance

• Prefer `coroutineScope { ... }` / `supervisorScope { ... }`.
• In apps, use lifecycle scopes (`viewModelScope`, `lifecycleScope`).
• Avoid `GlobalScope` except for truly process-long tasks.

Common pitfalls

• Launching work in a scope you don’t control.
• Not handling cancellation (e.g., ignoring `isActive`).