> -----Original Message----- > From: Barry Leiba <[email protected]> > Sent: 15 June 2020 18:11 > To: Jim Schaad <[email protected]> > Cc: Rob Wilton (rwilton) <[email protected]>; The IESG <[email protected]>; > [email protected]; [email protected]; Matthew > Miller <[email protected]>; Benjamin Kaduk <[email protected]>; > [email protected] > Subject: Re: Robert Wilton's No Objection on draft-ietf-cose-rfc8152bis- > struct-10: (with COMMENT) > > >> You need a normative reference on the CDDL document regardless, given > >> that it is not possible to understand the entirety of > >> draft-ietf-cose-rfc8152bis-struct without reading the CDDL draft, > >> hence it is a normative reference. The only alternative would be to > >> strip all the CDDL text out of the document, and I really doubt that > >> you would want to do that. > >> > >> Even if tools are just being used for validation then you still > >> require the CDDL to be right or otherwise the validation could give an > >> incorrect response. > >> > >> Hence, I would recommend that you bite the bullet, add the normative > >> reference, and make the CDDL text normative. Given that CDDL has > >> already been published as an RFC, I'm unsure why having a normative > >> reference to it would be a problem. > > > > [JLS] Section 1.4 is designed to avoid the need for CDDL to be > > normative. Also since the CDDL is explicitly normative I do not > > believe that it needs to be normative. > > I can't figure this out, Jim. Is there a "not" missing, or some such? [RW]
I suspect that Jim probably meant "CDDL is explicitly informative ...", but that doesn’t matter. I.e. just because the text is informative doesn’t make references to explain it informative (if required). Barry, we should try and complete the updated text for explaining normative vs informative references ... > > In any case, I agree that the CDDL reference can be informative. 8152 > predated the CDDL spec and was understandable with only an informative > reference to it, and I don't think 8152bis has changed in that regard. [RW] Actually, having looked at the draft again, I agree. I hadn't realized that this document aims to describe/document enough of the CDDL syntax that it doesn't require a normative reference. > > > I am still under the impression that one needs to look very closely at > > STD downrefs as well as Standards Track downrefs is that not true? > > An Internet Standard that cites a Proposed Standard normatively is > making a downref, if that's what you're asking. That said, it's not a > big problem to accept that -- we would just have to run another > two-week last call to approve the downref. But, as I said above, I > don't think that's necessary. > > Rob, can you give specific examples of things in 8152bis that can't be > properly understood without normative reference to RFC 8610? [RW] I've checked again, no I can't. Informative is fine. Regards, Rob > > Barry _______________________________________________ COSE mailing list [email protected] https://www.ietf.org/mailman/listinfo/cose
