Re: [netmod] draft-bjorklund-netmod-structural-mount / possible simplification

[t.petch <[e-mail]>]

----- Original Message -----
From: "STUART VENTERS" <[e-mail]>
To: "'Juergen Schoenwaelder'" <[e-mail]>
Cc: <[e-mail]>
Sent: Wednesday, March 16, 2016 2:58 PM


> Interesting, this highlights a concern I have with Yang.
> It's highlevel goals and progress are great, but at a detail level,
>    instead of using one general purpose language construct per
feature, it seems to be evolving to use many.
> For a new language evolving in an old problem area, this doesn't seem
right.
>
> RFC6020bis-11 section 7.15 gives some clue as to the reasoning behind
this.
> "The difference between an action and an rpc is that an action is tied
>    to a node in the datastore, whereas an rpc is not."
>
> To me, this feels like a protocol or implementation issue is causing
an unnecessary language addition.
>
> I wonder if an alternative strategy could be to say that you can use
an rpc in a node, but when you do the name of the rpc on the wire
becomes some combination of the node name and the rpc name?

Stuart

Picking up on your earlier e-mail, I agree with your thoughts but accept
Juergen's point that we should have got involved a year or two ago.

What I think would be valuable now is more guidance on when to choose
which option of the several available, like RFC6087 only more so.  I
think of this whenever I see a discussion on this list about e.g.
identities, enumerations, derived types, ...references and such like.
There are good reasons why one option is better than another but I
struggle to understand them and rarely remember when I am looking at one
of the dozens of YANG models that come out each month.

Tom Petch
>
>
> -----Original Message-----
> From: Juergen Schoenwaelder
[mailto:[e-mail]]
> Sent: Wednesday, March 16, 2016 9:08 AM
> To: t. petch
> Cc: STUART VENTERS; [e-mail]; Martin Bjorklund
> Subject: Re: [netmod] draft-bjorklund-netmod-structural-mount /
possible simplification
>
> On Wed, Mar 16, 2016 at 12:26:17PM +0000, t. petch wrote:
> > ----- Original Message -----
> > From: "Juergen Schoenwaelder" <[e-mail]>
> > Sent: Friday, March 04, 2016 4:52 PM
> >
> >
> > > On Fri, Mar 04, 2016 at 04:25:37PM +0000, STUART VENTERS wrote:
> > > >
> > > > 2)      Allow the 'rpc' and 'notification' nouns to be used in
other
> > places in the schema tree besides at the top module level.
> > >
> > > This is already part of YANG 1.1.
> >
> > Looking at rfc6020bis-11 s.14, I see 'notification-stmt' appearing
in
> > many places so indeed it is allowed in other places but only see
> > 'rpc-stmt' appear in 'body-stmts' .  Which, if I understand aright,
> > means that 'rpc-stmt' can still only appear at the top level.
> >
>
> Yep. But there is a new action-stmt and in section 1.1 it says:
>
>    o  Added a new statement "action" that is used to define operations
>       tied to data nodes.
>
> /js
>
> --
> Juergen Schoenwaelder           Jacobs University Bremen gGmbH
> Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
> Fax:   [phone]         <http://www.jacobs-university.de/>
>
> _______________________________________________
> netmod mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netmod