[NNTP] LIST EXTENSIONS and an NNTPv2 capability

[Clive D.W. Feather]

Ken Murchison said:
>> Not that I would necessarily go with his full proposal (perhaps a cut down
>> version). It might have been a good thing to consider earlier on, but not
>> at this late stage.

The whole matter never struck me until Russ's post the other day.

> Just out of curiosity, are you and Clive saying that a simple NNTPv2
> capability does not solve the problem at hand, or that it doesn't take
> things far enough?

The latter, I think.

I should have realized that what I was proposing was too big a chunk for
people to swallow in one go. Let me try again, explaining what I am
thinking.

(1) Do we need a way to determine that a server conforms to the new core
specification?

I think we all agree that the answer to this one is "yes". Let's not worry
about how for the moment.

(2) Do we need a way to determine that a server conforms to the new
specification for extension X, where we are changing that specification?

I think we do.

(3) Is it sufficient to say that servers conforming to the new core
specification must support the new specification for extension X?

No, it isn't.

There is a specific set of extensions - AUTHINFO, STARTTLS, STREAMING,
LISTGROUPS, HDR, OVER - where the new specification is being published
at the same time as the new core specification. It is therefore very
tempting to say that these must be kept in sync.

However, this doesn't suffice. Firstly because any update to any other
extension is not going to be published at the right time. And secondly
because extension definitions may themselves change.

Are you with me so far?

(4) Why not just have a new extension-label ever time an extension changes?

Well, we could do that, but let's consider the sitation a little way down
the line with, say, two minor changes having been issued to the definition
of OVER. Do we really want the LIST EXTENSIONS output to have to say:

    OVER
    OVERv1.1
    OVERv1.2

just to cope with the clients that haven't yet been updated? I don't think
so.

So this implies that we need some kind of versioning arrangement for
extensions, just as for the core specification.

Yes?

====

I have to stop writing as I need to be elsewhere. Can people please think
on what I've written so far and comment, and I'll try to continue later
today.

-- 
Clive D.W. Feather  | Work:  <clive at demon.net>   | Tel:    +44 20 8495 6138
Internet Expert     | Home:  <clive at davros.org>  | Fax:    +44 870 051 9937
Demon Internet      | WWW: http://www.davros.org | Mobile: +44 7973 377646
Thus plc            |                            |