Re: minutes for NETCONF WG interim meeting (09/03)

[Andy Bierman <abierman@cisco.com>]

At 09:53 AM 10/16/2003, Juergen Schoenwaelder wrote:

>>>>>> Andy Bierman writes:
>
>Andy> I think the key point is the need for parsing the identifier.
>Andy> We don't need to parse OIDs.  They are good as global
>Andy> identifiers via decentralized naming authorities.  URIs would
>Andy> also serve that purpose well.  But OIDs don't have version
>Andy> numbers.  The group at the interim felt it would be better to
>Andy> provide a distinct version number rather than create a strict
>Andy> URI format that every vendor must use for their own schema and
>Andy> capability definitions.
>
>Sorry, but we are talking about capabilities here. My implementation
>either understand a capability or not. So if a certain capability
>needs a revision so that the semantics change, you assign a new
>identifier. My implementation either understands this new identifier
>or not. In other words, what benefit do I get from being able to
>extract the version number in the first place? And if an agent
>supports multiple versions of the same capability, then it should
>announce multiple identifiers. Why do we have to make this more
>complex?

Sooner or later the concept of change control rules
will come up.  If there are any expectations that
version 1.1 of a capability implies any backward
compatibility with version 1.0, then it helps to
identify the version of the capability.  BTW,
these capability identifiers are also going to
be used for schemas, and we will definitely need
change control rules which include backward
compatibility expectations for different versions
of a data model.


>/js

Andy


--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>