[NNTP] Article number wording

[Ade Lovett]

Matthias Andree wrote:
> On Tue, 19 Jul 2005, Ade Lovett wrote:
>>Huh?  ARTICLE|HEAD|BODY|STAT <msgid> must be a figment of my imagination
>>then.
> 
> 
> They are part of the protocol and I'd say an indispensable one, but not
> in wide use by clients as their primary access method. The primary
> access method is GROUP <name> then XOVER or similar, then
> {ARTICLE|HEAD|BODY|STAT} <number>.

Maybe not in "true" Usenet clients (ie: text), but that's more due to
ancient history with certain NEWNEWS implementations being extremely
slow, resulting in a dramatic decline in its use (which is a shame,
since it is the perfect tool for a client operating across multiple news
servers, and removes the issues associated with a server failing
catastrophically, and resetting article numbers).

However, regardless of whether you view binaries-over-Usenet as "abuse",
this is the real world we're talking about, and it does happen.  A lot.
 A particular usage of Usenet cannot be sidelined on the basis of a
purist view, we have to be realists.  Usenet *is* used for binary
transmission, and there are certainly multiple-server clients out there
that use message ID as the primary access method.  Look up NZB files for
a case in point.

As and when a concrete situation is presented, with associated wording,
that makes sense, I'll certainly evaluate it on its merits, see this
mailing list passim for examples of that.

Right now, I'm not sure that the proposed wording does actually improve
the situation to the point where updates would have to take place,
possibly resulting in the entire document having to go back through the
whole IESG process (Russ?  How much flexibility do we have before the
document is changed to an extent that it would require this?).

-aDe