[NGO] modules vs. namespaces

[Andy Bierman <ietf at andybierman.com>]

Hi,

This thread on module names and version IDs demonstrates
that we are finally working on NETCONF content (!).

The NETCONF protocol just recognizes content as an XML blob,
consisting of a bunch of QNames that match data model nodes.
There is no concept of a 'module' within the protocol.

Now maybe we are finally moving beyond the 'XML blob', and a
NETCONF database can be managed as a collection of conceptual
modules.  Obviously, we have to allow for non-YANG modules,
but from a standards-POV, these will remain just QNames in the XML blob.

Instead of ad-hoc, highly variable support for specific
NETCONF operations, as allowed within the XML blob, and
documented in proprietary <appinfo> extensions,
humans and applications could have a consistent and precise
description of the configuration data, and the protocol operations
that must be supported on that data.


Andy

_______________________________________________
NGO mailing list
NGO at ietf.org
https://www.ietf.org/mailman/listinfo/ngo