Re: #1047 Path-identity: A proposal (perhaps)

[Frank Ellermann <nobody@xxxxxxxxxxxxxxxxx>]

Charles Lindsey wrote:

> So please can we be as brief as possioble in USEFOR, and
> point to USEPRO for the gory details

Hear, hear!  SCNR

> the term <path-identity> needs to include everything that
> may occur in the list, except the two keywords

...or the <tail-entry>.

> So it would be better as:
[...]

ACK.

> Here, now, is the text currently proposed for USEPRO.
[...]

>| To ensure that a <path-identity> provides a unique identity
>| for the news-server concerned, it SHOULD be one of:

How about "it MUST be" ?  Or even simply "it is" ?!?
[...]

>| The FQDN of a news-server is "mailable" if its
>| administrators can be reached by email using both of the
>| forms "usenet@<FQDN>" and "news@<FQDN>", in conformity
>| with [RFC 2142].

Maybe say 'forms "usenet@" and "news@" at the FQDN represented
by the <path-identity>, in conformity with [RFC 2142].'  Issue:
I'm not sure about using <FQDN> if that's no term in the ABNF.

>| "abuse@<FQDN>" MUST also be available

Dito, '"abuse@" address MUST' etc., if your <FQDN> doesn't fly.

>> Does this make sense?
> Not yet :-(

Getting much better, limiting ":" to IPv6 is a step forward.

Some of Bruce's points are not yet covered, MISMATCH!MISMATCH
etc., maybe we have to declare these beasts as special cases
of a path-delimiter, brain-storming:


 path-list = [FWS] path-identity 
             *( path-delimiter [FWS] path-identity [FWS] )
             path-delimiter tail-entry [FWS]

 path-delimiter = "!" / "!!" / "!MISMATCH!" / "!POSTED!"

No more path-keyword.  Does that solve Bruce's problem ?  Bye