#8 Separate input and output encoding

[Thomas Mack]

Just a wish: There is not too much sense in assuming
latin-1 output encoding, when the data can be in
UTF-8 or other encodings, as latin-1 or else is just
a subset of UTF-8. The data will always be resetted
to subset of characters of the original encoding, so
we cannot restore the characters to there original
values, I think.

So it would be good to have an input_encoding AND an
output_encoding selectable, to be able to use UFT-8
data on output.

It would be an alternative, to encode all data in
UTF-8 on output.