Re: [netmod] YANG next

[Ladislav Lhotka <[e-mail]>]

On Wed, 2019-07-24 at 22:37 +0200, Robert Varga wrote:
> On 24/07/2019 20:32, Andy Bierman wrote:
> > On Wed, Jul 24, 2019 at 10:28 AM Kent Watsen <[e-mail]
> > <mailto:[e-mail]>> wrote:
> > 
> > 
> > > > >     So you want to work on YANG 1.2, but just the parts you want to
> > > > >     change? ;-)
> > > >     I am actually fine with not doing any changes to YANG 1.1 at all,
> > > >     except perhaps
> > > >     bug fixes. This doesn't necessarily mean closing the NETMOD WG,
> > > >     it would IMO be
> > > >     immensely useful to rewrite the language specification and remove
> > > >     NETCONF- and
> > > >     XML-specific part.
> > > 
> > >     +1. There are plenty of ambiguities and NETCONF/XML pollution in the
> > >     spec. Having the specifications in a DAG would be immensely useful :)
> > 
> >     Agreed and I should've mentioned before that Martin said in Prague
> >     that he'd already started this effort, seeing it as a necessary
> >     pre-step before making other changes.  I'm unsure if the intention
> >     is to release this by itself as an RFC 7950 bis but, if looking for
> >     a minimal change, that might be it.  The next rung up would be to
> >     just add clarifications.  The next rung up from there would be to
> >     add only backwards-compatible changes (currently targeted by [1]). 
> >     The last rung being to also target NBC changes (there's no consensus
> >     to do this).
> > 
> > 
> > This WG sure likes to spend time refactoring documents.
> 
> Sorry, I am not watching very closely so I have not noticed.
> 
> I regard the NETCONF coupling in RFC6020/RFC7950 as technical debt,
> which when addressed will allow us to have better discussions about
> modeling intent vs. XML representation.

Absolutely. Also note that in RFC 7950 we even don't have a sound definition of
an instance data tree.

> 
> > Moving lots of text will create bugs and strong coupling, and only help
> > the standards purists.
> 
> Not necessarily. I agree with limiting the amount (and type) of text
> being moved and modifications done in between.
> 
> > It will be a lot of work for the WG and IESG to review such a massive
> > document split,
> > and in the end we have no improvement in YANG, just more RFCs to read.
> 
> No improvement in to YANG in terms of number of features, yes.
> 
> Even if it ends up being more documents, if they are smaller and more
> logically structured, they are more approachable.

Right. The base YANG spec should only define the language and rules for
validating instance data.

> 
> Reading RFC7950 (216 pages) is far from sufficient for understanding
> YANG, as you also need to understand NETCONF, which leads you down the
> RFC6241 rabbit hole.
> 
> If we can get a more approachable first document, lowering the entry
> barrier will be beneficial to people outside the WG(s).

I again fully agree. The entry barrier is unnecessarily high, which may not be
apparent to vintage members of the NETCONF/NETMOD gang.

Lada

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