It's not a question of whether it "can" or not sometimes. Say there's a performance issue in your code. In my experience, I have a much easier time figuring out

It's not a question of whether it "can" or not sometimes. Say there's a performance issue in your code. In my experience, I have a much easier time figuring out where the bottleneck is just by inspection with imperative code than with functional code (especially when concurrency is involved). Sometimes, transparency is more important than raw output.