ietf-822
[Top] [All Lists]

Re: UTF-8 over RFC 2047 (Re: Call for Usefor to recharter)

2003-01-11 15:22:39

Dave Crocker <dcrocker(_at_)brandenburg(_dot_)com> writes:
Erland> Why let it stop with 7 bits? Why not cram it into one bit, while 
you're
Erland> at it?

If this is is an example of the way discussions have gone on usefor, then
it's pretty clear why the technical work is in trouble.

7-bit netascii is a long-standing internet norm.  that's why.

If you were unimpressed by the argument, that was kind of the point.

It is quite clear from your comment that your knowledge on Unicode
is shallow. Believing that getting Unicode into a 7-bit world is
just to cram it into one less bit, is not very intelligent.

As already has been pointed out there existed the UTF-7 encoding for
this purpose. It has been withdrawn, because there too many problems
with it.

You're seven bit alternative appears to be RFC2047, but that is a
huge difference to a UTF-8 encoding. There is no editor, or any
other general-purpose tool that understands RFC2047. RFC2047 is
constrained to a much smaller realm.

Therefore going from raw UTF-8 to RFC2047 is a much huger leap than
stripping one more bit.

Erland> Incorrect. If you have an 8-bit editor that does not understand UTF-8,
Erland> you will see the text, but it will look "ugly".

it would be extremely helpful for this topic to benefit from attention being
paid to deployed, real-world standards-based installed base.

Yes it would. So why don't you do so, rather than referring to
theoretical 7-bit editors?

Erland> The same applies if you have a any-bit editor which does not 
understand
Erland> RFC2047.

Erland> But:
Erland> 1) UTF-8 looks less ugly than RFC2047 when incorrectly displayed.

Unfortunately, this effort is not a computer science theoretical exercise.

Good. Let's then aim at a solution that is practically useful, and which
have the least bad impacts when used with legacy software.

The theoretical ides about it could break some arcane software, I leave
to someone else.

--
Erland Sommarskog, Stockholm, sommar(_at_)algonet(_dot_)se