On Feb 5, 2007, at 5:25 PM, Neil Schemenauer wrote:

>  don't
> think I would make "utf-8" the default encoding but use "ascii",
> like the Python interpreter.

I'm trying to think of the practical impact of this choice.
The behavior would be different only if you have non-ascii,
not-utf-8 str instancess that you want to combine with (or use to
make new)
htmltext instances.  I imagine they would most likely come
from a database, or an external file.

If the programmer knows the encoding of this outside string,
then he/she will need to explicitly provide it to decode the
string.  Failure to do this will *always* result in an error if
the default is 'ascii', but the might go unnoticed if the default
encoding is 'utf-8'.  I think this is the case where 'ascii' is
better as a default value.

The 'utf-8' default has the advantage, however, in applications
that always store unicode instances using a utf-8 encoding.  If that
pattern is followed, then the programmer can stop working on encoding
puzzles.  I like that.

I also like having a default encoding for __str__() that can handle
non-ascii characters.  I like it that "print str(x)" does not ever raise
an encoding error when x is an h8 instance.