On Mon, Apr 30, 2018 at 05:57:34PM +0000, Kent Watsen wrote: > > I don't understand talk about abandoning this draft. There is no question > that it is needed (e.g., anima vouch, zerotouch, tail-f's "structure"), > and RFC 8040 is unsatisfactory because 1) it doesn't allow a top-level > 'choice' between two containers and 2) it requires drafts to reference > RFC 8040, even though the drafts may have nothing to do with RESTCONF. >
Re 1: RFC 8040 says: "It MUST contain data definition statements that result in exactly one container data node definition." So a choice may actually work as long as the result is exactly one container data node. OK, the wording in the RFC 8040 statement is not clear since 'result' and 'definition' do not line up (does 'result' mean the toplevel data node instances that are possible? In this case, 'definition' would be misleading). Re 2: It does not really matter whether you import the extension from RFC 8040 or some other module. Why is depending on A better than depending on B? The definition in RFC 8040 is already know by tools. I view the yang-data definition of RFC 8040 as a temporary solution, a proper solution should in my view be part of YANG 1.x. Since NMDA essentially binds all data tree definitions to datastores, the yang-data construct allows us to define data structures that are specifically not bound to any datastore, i.e., data structures that by design can't be operated on directly with NMDA NETCONF/RESTCONF. /js -- Juergen Schoenwaelder Jacobs University Bremen gGmbH Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany Fax: +49 421 200 3103 <https://www.jacobs-university.de/> _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod