ietf-nntp <0> and message IDs

[Russ Allbery]

Clive D W Feather <clive at demon.net> writes:

> However, with Russ's agreement, I could add a sub-section on the
> relationship between this document and Usefor, which would spell out how
> our requirements would be met in that context.

Something like this is probably not a bad idea, although personally I'd
rather see it as an appendix.  Basically, I think it wouldn't hurt to say
somewhere that this standard doesn't tell you everything you need to know
to write Usenet software that you need to read RFC 1036.

Also in that category are such things as warning people that a lot of
existing clients expect servers to follow RFC 2980 and will use XOVER
instead of OVER, etc.

None of that's normative; it's more just a warning to implementors of
things they may not realize if they read this standard in isolation.

>> Note also that INN for example suggests a message ID in the 340 response
>> to the POST command:
>> 
>> 340 Ok, recommended ID <b6rttc$jm3$1 at hermes.example.org>
>> 
>> The protocol might suggest that a client MAY use the suggested
>> Message-ID,

> If that response were meant to be meaningful, it would not conform to
> our standard.

What I think we're going to want down the road is an extension that, if
advertised, indicates that the textual part of both the direct response to
POST and the response after the article has been sent contain, at first, a
recommended message ID and then the message ID that the article was
actually posted with.  (The above free-form sort of way of giving a
message ID is obviously not what you really want.)

But we can talk more about that after the base spec is out.

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