Re: [DNSOP] Call for Adoptions: draft-lhotka-dnsop-iana-class-type-yang

[Paul Wouters <[e-mail]>]

On Thu, 10 Oct 2019, Ladislav Lhotka wrote:

> They should not actually be reading the RFC but get the latest revision of the module from this page:
>
> https://www.iana.org/assignments/yang-parameters/yang-parameters.xhtml

You are asking for text to go into an RFC, which you then say they
(implementors) should not read. Clearly then the text should not go
into the RFC.

> I do admit that this may not be clear to everybody. What we are really missing is a well-known and authoritative repository of YANG modules, at least those produced by IETF and managed by IANA.

Right. This is why I was hoping the IESG would have made some progress
on this, as they said they were working on "this".

> Would it help to reclassify the RFC as historic as soon as IANA takes over the responsibility for such a module?

I doubt it. Since developers that need to implement some RFC number have
no idea about the historic status. They tend to not even know the
difference between experimental, standard, informational.

>> I think this claim is both an unquantifed appeal to authority and unfair. The
>> technical debt of doing this happens 10+ years from now, when people are
>> implementing obsolete record types because a modern DNS Yang RFC references
>> these record types, and might be missing new record types because the
>> modern DNS Yang RFC did not yet have this new record type listed.
>
> I really don't understand. What DNS Yang RFC are you referring to? I am not aware of any.

Your draft.

> We are open to all reasonable suggestions.

My suggestion was a link the proper IANA registries, which _are_ updated
by other RFCs to place things into obsolete/deprecated and receive new
entries based on other new RFCs.

As you said the implementors need to go to the IANA/YANG module place, a
link to that would be more useful than including a current snapshot of
the IANA registries.

Paul