Guard clause vs Control Structures in #Elixir?

.md

1. Using guard

defmodule MyList do
  def max(list), do: _max(list, nil)

  defp _max([], max_value), do: max_value
  defp _max([head | tail], nil), do: _max(tail, head)

  # using guard
  defp _max([head | tail], max_value) when max_value < head, do: _max(tail, head)
  defp _max([head | tail], max_value), do: _max(tail, max_value)
end

2. Using cond control structures

defmodule MyList do
  def max(list), do: _max(list, nil)

  defp _max([], max_value), do: max_value
  defp _max([head | tail], nil), do: _max(tail, head)

  # using cond
  defp _max([head | tail], max_value) do
    cond do
      max_value < head -> _max(tail, head)
      true -> _max(tail, max_value)
    end
  end
end

Guard clauses happen at compile-time and can yield an optimization that cond cannot. However, guards are limited in what they can do and access, being compile-time. So cond has more flexibility.

@bcardarella Guard clauses provide performance vs cond provides flexibility, sound like go with guard clauses first unless it doesn't provide flexibility in the scenarios that are trying to solve.