Re: [netmod] WGLC: draft-ietf-netmod-geo-location-04

[tom petch <[e-mail]>]

________________________________________
From: Christian Hopps <[e-mail]>
Sent: 28 March 2020 13:39
> On Mar 28, 2020, at 8:36 AM, tom petch <[e-mail]> wrote:
>
>> A wholesale lack of YANG reference clauses; perhaps half a dozen needed
> I can see 2 places I might could put these, in the "astronomical-body" leaf that references the IAU and in the "geodetic-system" for the default value. We are creating an IANA registry for the values in geodetic-system though so perhaps you are asking for an IANA reference instead? I don't see 4 more obvious places for external references, could you help point them out?
> <tp> A good starting point is any reference in the body of the I-D should be in the YANG module too in a reference clause, such as www.iau.org, IS6709, WGS84 and may be more than once for different description clauses.  velocity could include the formula or refer back to RFCXXXX perhaps timestamp too.  RFC8344 has enough (but not too many IMHO).

I'll go through the doc, and see what else I might could add. A couple points in the meantime though...

Some of these references (non-normative) in the document are for the comparisons to other work, and they point at other outside normative definitions for similar objects. I think it would be wrong to put references in the YANG module that point away from the definitive work (this document) and towards some other normative standard.

For velocity, referring back to the RFC that defines the module within sub-statements of the module seems rather redundant. The reference at the top of the module is the default reference for the module, if one starts adding the same reference to module sub-statements where does it end?

<tp>
Christian
Think of an operator trouble shooting a velocity-related incident.  The YANG module is readily available and the description clause and any other part of the leaf definition is on the screen.  If that is all, and it is not enough, then the operator has to 
- pull up the entire YANG module
- scroll to the top and find the RFC  from which the module was extracted
- pull up the RFC and search it in its entirety
Contrast that with a YANG reference clause that says RFC XXXX s.2.3 which the OAM system may already have as a URL.
Likewise if the authority for syntax, semantics etc is a third party such as IAU, then better the reference clause takes the user straight there.
I think that the description clauses are good, often they are too thin, but sometimes more detail is wanted and that is where the YANG reference clause comes to the fore.
YANG modules need to be user-friendly when stand-alone.
Tom Petch




Other than that, I'll comb through again to look for normative references that aren't yet in the module, and also address your other concerns in a follow up version of the document.

Thanks for the review and comments!

Chris.