From: Russ Allbery (rra@stanford.edu)
Date: Tue Jan 09 2001 - 22:41:42 CST
Charles Lindsey <chl@clw.cs.man.ac.uk> writes:
> NOTE: The use of ".invalid" is to provide an aid to mail systems so that
> addresses deliberately intended to be malformed can be identified and
> delivery aborted. User agents MUST identify such addresses and require
> the user to alter the address when attempting a personal email reply.
> Injecting agents that have authentication information MAY choose to
> enforce the From-content based on the poster's authenticated identity.
I don't think putting conformance requirements in a note is a good idea,
and furthermore this is an interface issue. s/MUST/Ought/. No
interoperability problem is created by not treading .invalid specially.
> [Y1 it is my intention to adopt that wording.]
Agreed.
> [Y2 I hope nobody want to put any Oughts in there.]
> [Y3 I hope nobody want to put any Oughts in there.]
> [Y4 I hope nobody want to put any Oughts in there.]
> [Y5 Ought was SHOULD]
> [Y6 Ought was SHOULD. I hope nobody wants to change any of the other
> SHOULDs.]
I agree with all of these.
> [Y7 The above paragraph is to be rewritten as follows, now that we have
> the Injector-Info header.]
> Often this field will be the only entry in the region (known as the
> pre-injection region) after the '%', although there may be entries
> corresponding to machines traversed between the posting agent and the
> injecting agent proper. In particular, injecting agents that receive
> articles from many sources MAY include the identity of the source
> machine connecting to do the injection, (or other information to
> establish the circumstances of the injection) and SHOULD do so if no
> Injector-Info header (6.19) has been added.
I think we should strike the clause after the "and", including the SHOULD.
If the intention is to require trace information at the SHOULD level, that
should be handled in 6.19 with the Injector-Info header. I don't think
it's a good idea to try to standardize and require more than one place to
put this information, particularly when neither of them are existing
practice.
> Any such inclusion SHOULD NOT conflict with any genuine site
> identifier.
This seems like a very difficult requirement to fulfill if the above
recommendation is followed; how is the injecting agent supposed to know
whether there's also a news server running on the same host that uses that
as its Path entry?
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>