RE: [regrep] V3 Proposal: Namespace Manager Function

["CHIUSANO, Joseph" <JCHIUSANO@lmi.org>]

Title: RE: [regrep] V3 Proposal: Namespace Manager Function

I don't believe that having a namespace association at such a high level will cover the cases in which more than one namespace is referenced in a schema or XML document (I say documents because it is possible that someone can specify a local namespace in an XML document, thus overriding the target namespace).  Rather, it would cause a single namespace to be associated with all constructs in a schema or XML document, which many times will not be the case as multiple namespaces may be used.

Joe

> **************************************************************************
>   Joseph M. Chiusano
>   Logistics Management Institute
>   2000 Corporate Ridge
>   McLean, VA 22102
>   Email: jchiusano@lmi.org
>   Tel: 571.633.7722
> **************************************************************************
>

-----Original Message-----
From: David RR Webber - XMLGlobal [mailto:Gnosis_@compuserve.com]
Sent: Tuesday, January 29, 2002 11:21 AM
To: Farrukh Najmi
Cc: 'regrep@lists.oasis-open.org'; CHIUSANO, Joseph
Subject: Re: [regrep] V3 Proposal: Namespace Manager Function

Message text written by Farrukh Najmi
>
David,

I think, what Joe is looking for is ability to query based on the content
and
not the metadata. These content based queries would key off things like
namespace declarations and qualifiers within qualified names. This cannot
*ever* be done with Filter or SQL queries without support for content based
queries. Content based queries would handle this quite well actually.

<<<<

Farrukh,

We know everyone wants that ultimate capability.   What I'm saying is that
if you just add the namespace URI into the Classification metadata right
now
as a Note - then you can have this ability.   While this is not optimal -
its less
of a pandoras box than trying to write code to scan schemas and things
looking
for namespaces and then having the software making assumptions.

Since the namespace is often closely linked to "ownership" this should
work well with the registeredobject header metadata also as a "cue" -
as I'd previously noted - to make the dataentry easy for the users.

The key here is people being able to do very useful things right now
with V1.0 of Registry - without them thinking there's major holes we've
not filled or provided for.

I think his use case can be address significantly using V1.0 Registry,
and then we'll add extra functionality later as the icing on the cake.

Thanks, DW.