From: Andrew Gierth (andrew@erlenstar.demon.co.uk)
Date: Fri Jun 23 2000 - 11:47:01 CDT
>>>>> "Ralph" == Ralph Babel <rbabel@babylon.pfm-mainz.de> writes:
Ralph> Russ, I think that you - and many others on this list - are
Ralph> missing the point of Dirk's message. While we probably all
Ralph> agree that short IDs are desirable and that the limits need to
Ralph> be _documented_ somehow, these limits are _not_ an intrinsic
Ralph> part of the netnews message _format_
If we define a limit, then it _is_ part of the netnews message format.
Message-ids are sufficiently important and sufficiently frequently
used in different contexts that to require systems to cope with
arbitrary length ids would be both excessively burdensome and
completely unnecessary.
Within our standard the message-id has one purpose only - to provide a
globally unique identifier for the message. This does not require
arbitrary lengths, even on non-NNTP non-Usenet systems.
Ralph> Take, for instance, the maximum length of a "line" in the
Ralph> message body. Even though there's a well-known NNTP limit,
not so - NNTP has a limit on the length of _commands_ and _responses_,
not on the length of data lines.
-- Andrew.