Because it allows you to separate strings that use bytes from strings that use Unicode already in Python 2, and to continue that explicit difference when you get to Python 3 as well. I should add that ...