ietf-nntp Issues List version 3

Stan Barber sob at owlman.academ.com
Sat Nov 29 20:07:49 PST 1997


Topics for Discussion concerning "draft-ietf-nntpext-base-03.txt"

Stan Barber, Academ Consulting Services
Co-Chair, NNTPEXT WG Group
$Id: issues.txt,v 1.3 1997/11/30 04:07:26 sob Exp sob $

Things I know I need to put in the draft, so we don't need to discuss them
here:
	examples of how commands are used
	table of contents

Things that are issues:

Compatibility with RFC977 --
	SLAVE is not in the current draft. Do we add it back? If so,
	what is our rational for doing that? Do we now need a VERSION
	command so clients can tell the difference?

	What about using wildmat in various LIST commands and others?
	In RFC977, there was only a * and a ! when any matching
	on group names was permitted at all. Wildmat does not have 
	a ! in it? Should we add one for compatibility?

	What about wildmat on distributions? RFC977 didn't permit that
	at all. 

	Should NEWNEWS be included or not? Some argue that having it
	will discourage folks from using other means.

UTF-8 and wildmat
	The current draft offers one way to modify wildmat to accomodate
	UTF-8. Is that the way we should go? If not, what are the counter
	proposals? 

UTF-8, keyword and verbs
	The current draft advocates UTF-8 only be used for arguments
	in commands (not keywords or verbs). Is this the right thing to do?
	If not, what is the right thing to do? Is the current draft clear on
	this? If not, suggest some alternate text to make this clearer.

UTF-8 and responses
	In the responses where a fixed format is required (like ARTICLE,
	BODY, HEAD, and STAT), should the default be US-ASCII or UTF-8?
	[Right now, I think this is not a problem and perhaps we should
	just defer it until message-ids starting having 8 bit contents.]
	What about other responses? 
	[Note: I am only talking about the rest of the line following the 
	response code here, not the contents that might follow after that 
	in a multi-line response.]

BODY and MIME (a little pun there)
	The current BODY command does not provision including the MIME type
	in a reponse. Should this be added or is it best to assume that
	clients will get this header using PAT or HEAD if they want it
	before they get the body? If we add it to body, how do we do that?
	Should it be part of the response?

	What about lenghts here? RFC977 didn't have any lenghts on these
	kinds of things. Someone suggested 988 octects. Any opinions on this?

ABNF?
	The ABNF is based on the one posted to the mailing list. Does it
	match the text as written? Do we need to make any other changes
	to it (other than corrections to sync it to text).


OVER 
	Should we escape tab and newline characters in the header data
	returned by using the OVER command? If so, exactly how should this
        escape mechanism be defined? Or, should we just do what is
        currently done and convert any single instance of white space to
	one US-ASCII space character?

Other stuff--
	Do we want a table of response codes? Who wants to make the first
	pass based on this document?

	What other clarifications need to be made to the extension process?


Did I miss anything important?



More information about the ietf-nntp mailing list