Re: [CCAMP] WG Last Call on draft-ietf-ccamp-flexi-grid-fwk-03 and call for sheperd

[Daniele Ceccarelli <[e-mail]>]

Hi Matt, Ramon,

Since I found no clear statements on the usage of RFC2119 language with respect to this situations, I had a look at existing framework and requirement RFCs trying to find a common WoW. My take is:



-          Framework is always informational

-          Requirements are always informational

-          RFC2119 language is not homogeneous. Sometimes capital letters are used and sometimes not.

My preference is to use capital letters only when protocol behavior is defined, not when requirements for the design of the protocol are defined (this is in line with e.g. RFC7062 and RFC6163).
As I said this is just a preference, but if there is no reasonable objection I would suggest not to use any capital letter in the fwk+req document.

Thanks
Daniele

From: CCAMP [mailto:[e-mail]] On Behalf Of Matt Hartley (mhartley)
Sent: mercoledì 20 maggio 2015 19:12
To: Ramon Casellas; [e-mail]
Subject: Re: [CCAMP] WG Last Call on draft-ietf-ccamp-flexi-grid-fwk-03 and call for sheperd

Thanks, Ramon.

Hopefully the chairs and other knowledgeable folks can chime in on the 2119 language vs informational stuff.

Apologies on the requirements boilerplate – I just didn’t scroll down far enough :)

Cheers

Matt

Dear Matt,

Please see inline

El 20/05/2015 a las 17:07, Matt Hartley (mhartley) escribió:
Ramon,

Is this doc informational or standards-track? It claims to be informational, but it has RFC 2119 language in it.
It is a good question, which has been popping frequently during the lifetime of the draft. The last indication I received (internally) was that it should be informational, and I changed it. I guess the problem is that as a fwk document is mainly informational, but at some point in the distant past, it was decided to include also control plane requirements (I guess those are the sections that use RFC 2119 language). Maybe wiser people can comment on this, but it was my (admittedly limited) understanding that those are non-exclusive (i.e., RFCs may be informational but yet define requirements and use RFC2119 language to indicate requirement levels, adding the appropriate boilerplate).


Also, it’s missing the standard boilerplate for this stuff:

Uhm, afaik, it is not (I just checked both in -03 and -04 and the boilerplate is there $2.1). The change standards track -> informational happened in -04.

In any case, no objection either way. I will proceed with whatever option is more appropriate



Also: you currently have six main authors, and my understanding is that the RFC Editor’s guidelines are that there should be no more than five….
Thanks for pointing this out, we were told of this not so long ago,...
I would like to ask whether five is a rigid limit and whether exceptions may be granted (consider this a not-so-subtle request for exception...), given that this draft was the result of merging several drafts -- which explains the amazingly long list of contributors -- and authorship was a quite hard thing to agree on. OTOH I guess that if this policy really needs to be enforced, we will have to find a way (hopefully not involving dice)

Thanks for your feedback
Ramon