[NNTP] Additions to LIST commands
antti-juhani at kaijanaho.fi
Sat Nov 28 02:27:01 PST 2009
On Sat, Nov 28, 2009 at 11:12:00AM +0100, Julien ÉLIE wrote:
> Hi Antti-Juhani,
> >>Unless readers have a need for that information? Please speak up if
> >>that's the case. Thanks,
> >I suspect any group suddenly changing from y to x is going to cause questions,
> >which the server admin will have to answer in some out-of-band way. Better to
> >provide the information in-band (assuming clients start telling users that "r"
> >happened and what it means).
> OK, I understand your point.
It occurs to me now that much more useful for that purpose is just to have a
customized message along with the POST rejection. LIST ACTIVE is not going to
be commonly checked, unless there's some what I'm forgetting for the client to
request updates in the style of NEWGROUPS.
Thus, my argument for "r" vanishes.
Perhaps there should be a way for a server to provide in-band a group-specific
motd message, but this can be simulated fairly easily by having the admin post
a local-only message to the newsgroup.
And in any case, I'm a bit wary about adding unimplemented features in
standards, even though a Proposed Standard is allowed them.
> Other currently out-of-band information is the list of hierarchies for which
> control articles are (supposed to be) honoured, the list of NoCeM issuers whose
> notices are (supposed to be) honoured...
> But maybe no one cares about having these lists in-band?
> LIST EXPIRE seems to have the same interest as LIST REMOVALS, doesn't it?
These differ from removals in that the data is stable - you look it up once,
and you're likely not going to get into trouble for using outdated info. In
contrast, a delayed removal (y->x before actual removal) is going to be a
sudden and (in general) unforeseen event, which, when it happens, will raise
questions. But my comment above applies as to what's the best way of providing
Antti-Juhani Kaijanaho, Jyväskylä, Finland
More information about the ietf-nntp