IMHO, I am no expert at all in which encoding is most commonly used in practice, but I cannot even imagine myself ever wanting to choose ASCII first. Since we might change the readme right now, could we make UTF8 the default and mention ASCII as another possible option instead?
I have one old and unimportant database that I created with ASCII when I used postgresql for the first time after following the instructions as given. I regret the choice, not that it really ever mattered for that database. Thanks, Chris Bennett