[ietf-nntp] Draft 21

[Ken Murchison]

Clive D.W. Feather wrote:
> Okay, there's a pre-release of the draft at
> <http://www.davros.org/nntp-texts/draft21.pre-1.txt> and
> <http://www.davros.org/nntp-texts/draft21.pre-1.html>.
> 
> The former has diff-marks in it; they were hand-generated, so won't be
> repeated in future.

Thanks Clive.  I have a few comments/questions:

- How should LIST NEWSGROUPS handle a group which has no decsription? 
The language in the draft leads me to believe that if an empty response 
is given, that the client might infer that there are no groups 
available.  Is this correct?  I assumed that LIST [ACTIVE] is used to 
get a list of available groups and LIST NEWSGROUPS was just 
informational (it returns descriptions for those groups that have them). 
   Or does LIST NEWSGROUPS have to return a line (with possible empty 
description) for every groups that is returned by LIST ACTIVE?

- How should the HDR ALL capability interact with the new LIST HEADER 
MSGID?  The reason why I proposed HDR ALL was so that the server didn't 
have to list every possible header that it can fetch.  Now that we are 
allowing HDR to return different [sub]sets of header/metadata depending 
on the form of HDR, the simplicity of HDR ALL has gone away.  For 
example, how do we handle a server which can return all headers when 
fetched by msg#, but can only return metadata when fetched by msgid (or 
vice-versa)?  Using HDR ALL will be misleading, since is only applies to 
half of the cases.  Perhaps we should get rid of HDR ALL and replace it 
with a different response code (214, 216?) for LIST HEADERS which means 
"all".

-- 
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp