ietf-nntp another draft released

[Brian Hernacki]

I'm a little concerned with the addition of the CHARSET verb to the base
protocol. 

Initially I thought this was a "good thing". It seemed nice that we were
taking into account internationlization issues. But after thinking about
it for a while and talking to folks who have worked on similar tasks in
other protocols, I've changed my mind.

I think CHARSET is the wrong direction to head in. By allowing clients
to specify charsets, we're going to end up splintering a space which
right now is unified. This would, in my opinion, probably make us worse
off than we are now. We'll have many different clients trying to use
many different charsets. I doubt many servers will implement all of the
charsets clients will use, so we'll just end up with a steaming pile of
uninteroperabilty.

What I'd like to see us do is just decide on a charset for NNTP to use
and put that in the base document. After looking at IMAP and talking to
folks at the Munich IETF, I'd suggest using UTF8 for this. Remember that
charset is not language, it's an encoding for languages. If we select a
good one, we should be able to use it for most, if not all languages.

--brian