Hannes,

> This boils down to the question if Latin1 or UTF8 should be the default
> text encoding. If one thinks backwards Latin1 is probably the choice
> whereas if we look forward UTF8 is surely to be preferred.

  Even if one thinks backwards, latin1 wouldn't be the choice.

  The argument Ned raised was not based on the text; is was on
unspecified data.

  In my opinion, those things should be byte streams and the code
should pay better attention on the text/binary distinction.

-- Yoshiki