Re: BCP97bis a process problem

[tom petch <[e-mail]>]

On 19/10/2021 21:46, Brian E Carpenter wrote:
> On 20-Oct-21 00:26, Michael Richardson wrote:
>>
>> Brian E Carpenter <[e-mail]> wrote:
>>>> I can think of two explanations.
>>>>
>>>> 1) the author is not familiar with the processes and thinks that I-D
>>>> Nits it ordering the author to reclassify the reference
>>>>
>>>> 2) the author is familiar with the processes and is trying to game the
>>>> system.
>>
>>> There's a possible 3rd explanation:
>>
>>> 3) the reference didn't need to be normative in the first place.
>>> I certainly saw that a number of times as a Gen-ART reviewer.
>>
>> How can a third party point this out, unless they can read the document?
>> The converse is worse: some informative reference that actually is normative,
>> but only implementers two years later recognize this.
>
> Well yes. That's exactly why, during my career as a Gen-ART reviewer, I tried
> to pay attention to this whole issue. But there is scope for human error here,
> so it's going to happen from time to time.

Brian

  Ah yes, I did not consider the third option.

To make this concrete, a recent I-D has a Normative Reference to a 
protocol RFC that was clearly Normative but was now being updated to a 
-bis.  I commented that a -bis existed and was well on its way to RFC so 
had the authors considered referencing that.  A revised I-D appeared 
with an Informative Reference to the -bis!  Mmmm; I suspect that some 
flag in the process pushed the authors into that downgrade.

A regular example is the data model with a reference to the information 
model that is Informational. To me it is obvious that information models 
should be Standards Track since the first reference to them will be 
Normative but WG Chairs do not see that :-(

Tom Petch

>      Brian
> .
>