[NNTP] Re: New NNTP drafts approaching IETF Last Call

Russ Allbery rra at stanford.edu
Thu Mar 24 10:54:46 PST 2005


Charles Lindsey <chl at clerew.man.ac.uk> writes:
> Russ Allbery <rra at stanford.edu> writes:

>> Please note that this is not, as I understand it, an accurate
>> representation of the IETF stance on stringprep.  If we were starting
>> from scratch today with NNTP as a new protocol, I don't think there
>> would even be any useful discussion about whether or not we would have
>> to deal with stringprep.  It would be a requirement.

> I am not convinced of that. It is certainly legitimate to use
> charset=utf-8 in the body of an article (or in a part of a multipart) -
> usually in conjunction with CTE 8bit.

Article bodies are a different issue.  Stringprep comes in to play with
command arguments, not with data transport.  I agree that we don't need
stringprep with articles.  (What we would likely be required to do is
state that articles must comply with MIME, were we starting from scratch
right now.)

> So any requirement for stringprep (or any other -prep) could only arise
> with texts generated by the server (e.g. in responses), and not in
> bodies of articles passing in either direction.

And in arguments to commands issued by clients.

> But does the IETF actually require normalization of texts that are only
> intended for human consumption?

Newsgroup names are not only intended for human consumption.  They are,
for example, used as arguments to commands issued by clients.  An obvious
analogy is to usernames, and SASL was indeed required to specify
stringprep.

> As I said yesterday, newsgroup-names are a matter for the format
> standards to sort out (in the current Usefor they are still US-ASCII).
> There is also the matter of the output of the LIST NEWSGROUPS command,
> but that again is a matter for Usefor or its successors (since it is
> closely tied with the checkgroups control message, which again is
> currently restricted to ASCII).

Note that NNTP does not restrict either of those to ASCII, which is why
this comes up.  Nor do I think we should.  NNTP is significantly in
advance of USEFOR here at the moment, which is not a bad thing, but it
does mean that we have to think about this issue and see what we need to
do to deal with it.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>



More information about the ietf-nntp mailing list