Re: [netmod] operational knobs

[Andy Bierman <andy at netconfcentral.com>]

Jon Saperia wrote:
> 
> On Nov 1, 2009, at 2:35 PM, Andy Bierman wrote:
> 
>> Jon Saperia wrote:
>>> Randy,
>>>
>>> Good points.
>>>
>>> In the context of the current conversation, this history argues for
>>> getting as much of NETCONF/YANG correct the first time.  I mean not only
>>> for the immediate application to which the protocol and language will be
>>> put, but also, where people want it to go.
>>>
>>
>> I disagree.
>> We have a lot of theory, but not 1 config=true
>> object in any I-D or RFC.  We have a description-stmt
>> for a reason.
>>
>> It is much easier for a WG to reach consensus
>> on advanced features if there is common experience
>> based on well-known data models in use.
> 
> I was not suggesting chrome polishing or advanced features.   Yesterday
> you asked about operational knobs and said current drafts do not support
> them.  My point is that, if as you suggest this is to be a requirement,
> that there is consensus  that it can be addressed incrementally, if not
> right now.

The consensus seems to be that 'action' (rpc-stmt) is
the correct choice, not a config=true leaf.

One question that keeps coming up: 'what is config?'
There seems to be consensus that config is not
a leaf that the server changes, once the client has set it.


> jon

Andy