Hi Robert, > On 13 Oct 2015, at 17:26, Robert Varga <[email protected]> wrote: > > On 10/07/2015 07:37 PM, Kent Watsen wrote: >> >> This is a notice to start a NETMOD WG last call for the document: >> >> Defining and Using Metadata with YANG >> http://tools.ietf.org/html/draft-ietf-netmod-yang-metadata-02 >> > > I have read the document, and coming in without the previous discussions, > md:annotation instances feel like aspects (from aspect-oriented programming) > being attached to pre-defined models to address a cross-cutting concern. > Unfortunately it seems it lacks an equivalent of pointcut specification, e.g. > a machine-readable definition where/when a particular annotation is valid. > > This makes annotations disconnected from the YANG metamodel, which is not > desirable for systems strictly based on the metamodel, as each instance of an > annotation will require hand-written code. > > My question is whether it would make sense to define some sort of (optional) > pointcut specification to be carried within the md:annotation statement (such > as an explicit list of data nodes, or a 'when' statement or similar)?
The easiest way would be to define "annotation" as a first-class data node. However: - YANG was deliberately designed to avoid using XML attributes in the same way as XML schema languages, - IMO this can't be done through an extension. With must or when it would be difficult to determine the context node. Lada > > That would allow us to bind some/most md:annotation instances automagically > to the metamodel, reducing the need to hand-code their semantics. > > Thanks, > Robert > _______________________________________________ > netmod mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/netmod -- Ladislav Lhotka, CZ.NIC Labs PGP Key ID: E74E8C0C _______________________________________________ netmod mailing list [email protected] https://www.ietf.org/mailman/listinfo/netmod
