RE: FW: I-D ACTION:draft-weijing-netconf-interface-00.txt

["Allen, Keith" <Keith_Allen@labs.sbc.com>]

Randy does a good job below of answering Andy's first question for me.  I am
not proposing that netconf use anything other than XML Schema to define data
models.  As Randy points out, though, as a data definition language the XML
schema standard doesn't really give us everything we need, so I believe we
will end up defining conventions on the use of XML schema to fill in the
gaps.  This is where the challenge that I spoke of comes in.  For example,
it's important that a data model define what information can be modified and
what can only be read.  The Enns draft defines one convention - that
read-only data be contained in one <state> element in a device's schema,
while draft-weijing proposes that it be identified with an attribute.  If
XML was an object definition language, it would probably define a standard
way of identifying what was readable and what was writable, and we wouldn't
have this problem to solve.


Keith Allen
SBC Labs
9505 Arboretum Blvd.
Austin, TX 78759
(512) 372-5741
keith_allen@labs.sbc.com
 


-----Original Message-----
From: Randy Presuhn [mailto:randy_presuhn@mindspring.com] 
Sent: Tuesday, July 08, 2003 11:58 AM
To: netconf@ops.ietf.org
Subject: Re: FW: I-D ACTION:draft-weijing-netconf-interface-00.txt

Hi -

> From: "Andy Bierman" <abierman@cisco.com>
> To: "Allen, Keith" <Keith_Allen@labs.sbc.com>
> Cc: "'Juergen Schoenwaelder'" <schoenw@ibr.cs.tu-bs.de>;
<netconf@ops.ietf.org>
> Sent: Tuesday, July 08, 2003 9:28 AM
> Subject: RE: FW: I-D ACTION:draft-weijing-netconf-interface-00.txt


...
> >The challenge is defining *objects* with a "data definition language," as
> >opposed to an object definition language.
>
> This point is not very clear at all.  Do you have any examples
> of 'objects' that cannot be defined with an XSD?  What are
> you proposing netconf uses for defining data models?  Note that
> this is somewhat outside the scope of the WG.  We will use
> XML Schema to define the netconf protocol syntax, but this
> does not mean that all data models will be defined that way.
> However, we can make use of valid XML constructs, and expect
> that any reasonable data modeling language will support them.
...

An object definition language has facilities for representing semantics
and relationships that go well beyond those of a data definition language.
Two examples:  ASN.1 vs GDMO, and ASN.1 vs SMI(ng).  I don't know
what Keith has in mind or whether he's proposing anything, but
an XSD would not be terribly well-suited to representing the semantic
constraints on the objects referred to by the arguments of a method
associated with an object.  But, as you say, all this practical stuff is
out of scope, so I'll shut up.

Randy

Randy



--
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/>

--
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/>