<focussing on datastores> ----- Original Message ----- From: "Ladislav Lhotka" <lhotka@xxxxxx> To: "tom p." <daedulus@xxxxxxxxxxxxx> Sent: Thursday, February 25, 2016 1:26 PM Tom, > On 25 Feb 2016, at 13:42, tom p. <daedulus@xxxxxxxxxxxxx> wrote: > > In the interests of clarity > > - datastores are not mentioned. These loom large in YANG and NETCONF > and, I think, have been misunderstood by those wishing to extend YANG in > various, new directions. Therefore I think that the I-D should say > something, even if it is that the concept of datastore is alien to the > envisaged uses of JSON (I could envisage a use where datastores do > apply, but it is probably an unrealistic use:-) I don't understand. This draft is about encoding a data tree in JSON under the assumption that the data tree is valid with respect to a YANG data model. How is this related to datastores? In particular, I don't think the concept of datastores is alien to it in any way (proofs exist to the contrary). <tp> It is the I-D that introduces datastores " The specification of YANG 1.1 data modelling language [I-D.ietf-netmod-rfc6020bis] defines only XML encoding of data trees, i.e., contents of configuration datastores, state data, input/output parameters of RPC operations or actions, and event notifications. The aim of this document is to define rules for encoding the same data as JavaScript Object Notation (JSON) text [RFC7159]." and goes on to give a definition of action and RPC operation but not of configuration datastore, state data or event notification. To me, that looks odd. The I-D tells me I could take rfc6020bis and replace every XMP snippet with JSON text and for that, I think I need a knowledge of datastores! I suggest adding those three missing definitions to section 2, nothing more. Tom Petch <snip>