[NGO] Decoupling Content definition from Modelling Language?

["Sharon Chisholm" <schishol at nortel.com>]

Hi

I think this list is still the right place for slightly off-topic
discussion of NETCONF related discussions.

Conceptually, the NETCONF architecture, similar to the SNMP
architecture, consists of
  - The protocol
  - The modelling language
  - The content

While in practice with SNMP, we did not change things up as much as we
could have (two languages,  three protocol versions, one content), this
did prove to be a useful construct.

One thing I've been pushing for, although from some hallways discussions
people seem to have completely misunderstood this, is to leverage this
separation to allow us to start defining the much needed NETCONF content
without waiting for the completion (RFCs published) of the NETMOD work.
So far this has taken two forms
  a) publishing in XML Schema (as Andy points out, our normative
language for now)
  b) and more recently publishing in both XML Schema and yang (without
being normative if it isn't done yet when published)

These seem to be working more or less, but I'm still sensing some
apprehension on moving forward with content standardization. From a
content perspective, it seems like syntactic sugar to me.  Are there
specific technical issues with this approach that I'm not aware of?

Sharon Chisholm
Nortel 
Ottawa, Ontario
Canada
_______________________________________________
NGO mailing list
NGO at ietf.org
https://www.ietf.org/mailman/listinfo/ngo