>>35
But nowadays everything is UTF-8 so it's quite reasonable to use it as a default encoding instead of plain old ASCII. Of course, internally it can just convert everything to UTF-32 (and many languages/implementations do) and be done with it but it shouldn't make me encode and decode everything when I'm using
THE STANDARD.
At the very least it should give me the ability to define a default IO encoding from which everything should be converted to/from.