Re: [netmod] WG Last Call: draft-ietf-netmod-schema-mount-07

Juergen Schoenwaelder <j.schoenwaelder@jacobs-university.de> writes:

> On Fri, Nov 10, 2017 at 04:39:36PM +0100, Ladislav Lhotka wrote:
>> >
>> > So what is the difference between "schema tree" and "schema"? Or to
>> > put it differently, what is "all associated semantics" that you are
>> > adding to a "schema tree" to obtain a "schema"? RFC 7950 says:
>> >
>> >    o  schema tree: The definition hierarchy specified within a module.
>> 
>> As Rob points out, this is probably incorrect, as schema tree should
>> involve multiple modules. It makes no sense to talk about a schema tree
>> of an augmenting module unless we also take into account the augmented
>> module.
>> 
>> Anyway, schema tree is really just the hierarchy, i.e. a tree of schema
>> nodes, whereas schema is the hierarchy with datatypes, semantic rules
>> and all that. We can validate instance data against the schema, not
>> against the schema tree. This is at least how I understand it.
>
> The current definition says "definition hierarchy specified within a
> module", perhaps this was intended to mean 'tree of schema nodes', but
> perhaps it means "the hierarchy with datatypes, semantic rules and all
> that". ;-) Looking at the definition of schema node, I see:
>
>   o schema node: A node in the schema tree.  One of action, container,
>   leaf, leaf-list, list, choice, case, rpc, input, output,
>   notification, anydata, and anyxml.
>
> With this, your interpretation makes sense and a better way to define
> scheme tree would then be:
>
>   o schema tree: The tree formed out of schema nodes of a module.
>
> The schema could be defined as
>
>   o schema: All definitions of a module including the schema tree.
>
> I think it is OK to scope these definitions to the notion of a module.
> If people talk about the schema (schema tree, schema nodes) of a
> collection of modules, it seems natural to assume that this means the
> union of the schemas (schema trees, schema nodes).

In many cases it is more useful to talk about the multi-module schema as
defined by YANG library and, potentially, schema mount. The
tree-diagrams draft assumes that tree diagrams will capture both the
parent and mounted hierarchy in the same diagram. The thing is that a
collection of modules often defines hierarchies that aren't mere
side-by-side unions of the individual module hierarches. And
specifically, it is somewhat difficult to talk about a hierarchy for a
module that consists only of augments, without having the context of the
module(s) augmented by it.

So I would suggest to define both "module schema/tree" and
"schema/tree", the latter meaning a multi-module schema as defined by
YANG library and schema mount specification.

Actually, we have another candidate for a more precise definition, namely
"data model". So one option could also be to use "schema/tree" for the
single-module, and "data model/tree" for the multi-module stuff.

Lada

>
> /js
>
> -- 
> Juergen Schoenwaelder           Jacobs University Bremen gGmbH
> Phone: +49 421 200 3587         Campus Ring 1 | 28759 Bremen | Germany
> Fax:   +49 421 200 3103         <http://www.jacobs-university.de/>
>
> _______________________________________________
> netmod mailing list
> netmod@ietf.org
> https://www.ietf.org/mailman/listinfo/netmod

-- 
Ladislav Lhotka
Head, CZ.NIC Labs
PGP Key ID: 0xB8F92B08A9F76C67