Lada,

On 3/8/2017 11:40 AM, Ladislav Lhotka wrote:
>> On 8 Mar 2017, at 16:05, Lou Berger <[email protected]> wrote:
>>
>> Martin, Juergen,
>>
>>
>> On March 7, 2017 8:08:26 PM Martin Bjorklund <[email protected]> wrote:
>>
>>> Juergen Schoenwaelder <[email protected]> wrote:
>>>> On Fri, Mar 03, 2017 at 04:41:44PM +0000, Kent Watsen wrote:
>>>>> All,
>>>>>
>>>>> Lou and I were discussing how it seems unnecessary that every draft
>>>>> has the same boilerplate text regarding how to interpret tree diagram
>>>>> notations.  It would be nice if drafts could instead just reference
>>>>> another draft that contains this information.  Does this make sense?
>>>>>
>>>>> Assuming we're interested in having such a reference, we could define
>>>>> a mini-RFC or, perhaps, leverage Section 3 of 6087bis (YANG Tree
>>>>> Diagrams).  Either way, we'd want/need to ensure the information
>>>>> is updated in a timely manner.
>>>>>
>>>>> Two reasons for why we may not want to pursue this are:
>>>>>  1) we can’t update the reference fast enough
>>>>>  2) drafts might add some proprietary annotations
>>>>>
>>>>> Is this worth pursuing at all?
>>>> This has been discussed before. The tree format that tools generate
>>>> has evolved a bit over time and the current setup allows to have some
>>>> evolution. The question is whether we have reached a state where the
>>>> evolution has come to standstill and we can nail a common tree format
>>>> down.
>> I don't see that as the question at all - the issue for me is needing to
>> parse each document to see if and how it differs from the norm and then
>> figuring out if the differences are (a) a bug, (b) limited to the
>> specific document, (c) something that is a basic change that should
>> impact tools (i.e., pyang) and other documents.
>>
>>> I don't think so.  For example, it was recently suggested that a
>>> notion for "mount-points" should be defined.
>>>
>> Yes, and it is our (Martin, Lada and my) conversation in that context
>> that prompted this discussion.
>>
>>> I don't think this is a big problem.
>> Again, I do see this as an issue worth solving and am appreciative that
>> 6087bis is available to easily provide a stable reference until such
>> time as an update/replacement is needed.
> If the format itself isn't stable, how can 6087bis (after it becomes an RFC) 
> provide a stable reference?

huh? - any RFC can be updated per normal process whenever appropriate. 

> I agree with Juergen and Martin and don't mind having the section about tree 
> symbols in each document that needs it.

I completely disagree - it means we constantly need to do diffs. 
There's a good reason to reference prior work when *not* changing
something which shows up, in this case in potentially 10s if not 100s of
drafts.  This allows *everyone* to immediately notice when something new
or unique is done.  (I'll have to find it, but I really loved the RFC
that used a well known term, but then redefined it for that one and only
document -- why do we want to allow this???)

What benefit comes from defining tree syntax by reference?

BTW if you/the WG prefers for the definition to be in its own document,
that would work too.

Lou

> Lada
>
>> Lou
>>
>>>
>>> /martin
>>> _______________________________________________
>>> netmod mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/netmod
>> _______________________________________________
>> netmod mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/netmod
> --
> Ladislav Lhotka, CZ.NIC Labs
> PGP Key ID: 0xB8F92B08A9F76C67
>
>
>
>
>
>

_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod

Reply via email to