This discussion may be conflating two issues:
(i) Does RFC text have to use RFC2119 terms to be normative?
RFC 8174 categorically states that text can still be normative without
using RFC 2119 terms.
(ii) Should standards track documents use RFC 2119 terms?
If 93% of recently published standards track RFCs make use of RFC 2119
terms then that seems like a strong consistency argument to use them
unless there is a good reason not to.
We've already agreed that the datastores draft will use RFC 2119 terms
where appropriate. For the model drafts, Benoit's suggestion to state
that the appendix is normative seems like an easy solution where it is
required.
If it gets discussed at Singapore then I would suggest doing so at a bar
rather than spending WG time on it ;-)
Cheers,
Rob
On 02/10/2017 12:58, Lou Berger wrote:
Juerge,
Understood. I think you made this clear in our previous discussion on
this topic, even though ~93% of the RFCs published in the last 5 years
use it. We certainly can discuss this with our AD, and if there's
sufficient interest in the WG even discuss it in Singapore. If others
are interested in face to face time for such a discussion, please let us
(all) know on the list.
Cheers,
Lou
On 10/2/2017 7:05 AM, Juergen Schoenwaelder wrote:
Lou,
the conclusion is that we add RFC 2119 here and there but I disagree
with the notion that normative text needs RFC 2119 language, i.e.,
that text that does not use RFC 2119 language is not normative. See
the pointers to the RFCs that I have provided. Now you want to make
this even a rule for all future WG docs so I strongly oppose to that.
/js
On Mon, Oct 02, 2017 at 06:39:35AM -0400, Lou Berger wrote:
Benoit,
I think this and related topic was closed with the conclusion of sticking
with 2119 language for normative text in current and future WG docs. We
certainly can add this sentence as well.
Lou
On October 2, 2017 5:11:20 AM Benoit Claise <[email protected]> wrote:
Dear all,
To avoid any confusion, just clearly mention it.
"This appendix is normative | informative"
No need to debate for hours on this.
Regards, Benoit
----- Original Message -----
From: "Lou Berger" <[email protected]>
Sent: Thursday, September 14, 2017 6:06 PM
On 9/14/2017 12:36 PM, t.petch wrote:
Appendices are Normative if they say that they are Normative. The
default is that they are not so say that they are and they are.
This is
well established practice.
Hi Tom,
My memory (I haven't checked recently) is there is nothing in or
defined process that says if an Appendix is normative or not. Other
SDOs certainly have formal definitions here. Within the IETF, my view
has been that if an appendix includes RFC2119 language, it is
normative. Actually, strictly speaking, any text in a Standards Track
RFC that doesn't include RFC2119 language is just informative.
Lou
Try RFC4910.
' This appendix is normative.'
and not a SHOULD or a MUST in sight.
Tom Petch
Lou
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
.
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
.
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
