Re: [I2nsf] Genart last call review of draft-ietf-i2nsf-nsf-monitoring-data-model-12

[t petch <[e-mail]>]

On 02/12/2021 03:22, Dale R. Worley wrote:
> Addressing what seems to be the difficult issue:
>
> t petch <[e-mail]> writes:
>> The trouble with the way that Last Call is organised is that the changes
>> suggested below will pull this I-D out of line with the others
>> potentially leading to contradiction and confusion.
>
> As I understand it, the only critique I presented of the overall
> organization is:
>>> But since the data model definition
>>> does not depend on the overall architecture, the document should be
>>> revised to either (1) remove unnecessary references to the overall
>>> architecture, (2) segregate them in ways that show they are not needed
>>> to understand the data model, or (3) carefully referenced back to the
>>> documents that define them.
>
> My intention was to suggest multiple ways that this draft could be
> adjusted so that it was clear to someone who hadn't already absorbed the
> gestalt of the whole document set.  There are likely other possibilities
> as well.
>
> It seems to me that the approach that would require the least change is
> (3), updating the terminology section to list all the terms that are
> imported from the other documents and provide references to their
> definitions.
>
> I would also give some consideration to whether there are
> definitions/descriptions of the data items that could be clearly
> understood without external references, but the current text requires
> that external context to be easily understood.  In such cases, it's
> usually best to provide a "context-free" defintion, and then add an
> explanatory sentence describing how that fits within the larger system.
>
> But I don't see any of this as pulling this draft away from the others.
> Is there an example that comes to mind?

Dale

I think that your comments are excellent as comments on a isolated I-D 
but might not be such an improvement for the set of six (or seven or 
eight) that I reference.  When these I-D first appeared, there was a 
debate on the WG list.  This is not Event-Condition-Action. Oh yes it 
is. Oh no .. (we are in the pantomine season).  I have seen a number of 
such divergences since.  Your comment that there are two different 
definitions of Event in this I-D reminds me of this.

If you add a terminology section to this, as I have seen other WG do, 
then you need to update all the other I-D with something similar IMHO 
and I have seen this take a lot of effort to get that consistent.  I had 
forgotten, perhaps never knew, that there is an expired I-D on I2NSF 
Terminology.  WIth hindsight, putting it all in there would simplify my 
task, but then might complicate one whose only interest is in - say - 
nsf-monitoring who would then need to reference two I-D instead of one.

My approach has been to focus on the YANG and make that coherent with 
the other I-D in the set.  I have mostly ignored the text.  Your review, 
to me, focusses on the text and the inconsistencies therein and not the 
YANG.  My concern is then that changing the text will introduce 
inconsistencies between text and YANG and between this I-D and the 
others in my set.

There is a similar issue with the YANG that there is much overlap within 
the set of YANG modules (which I have worked to make consistent).  WIth 
hindsight, a common module, two or three years ago, would have 
simplified my work but, as before, someone with only an interest in - 
say - nsf-monitoring would have yet another Normative Reference to 
become familiar with.

Tom Petch

> Dale
> .
>