Re: Finalised Glossary Definitions

Mark, Jean-Jacques

Here is my two cents.  Please note I am not claiming to be be  a
namespace person.
I am not in full agreement with Mark's comment.  A quick examination of
an RPC module implementation reveals that what might be considered a
handler is targeted.   Particularly, I am thinking of the URI of WSDL
file.

Regards, Marwan

Jean-Jacques Moreau wrote:
> 
> Mark,
> 
> Thanks for your answer.
> 
> Mark Nottingham wrote:
> 
> > To me, the namespace is a natural way to do this; it identifies the
> > Module (current definition) whose functionality is desired, and
> > installed handlers on the node will identify which functionalities
> > they implement, by the same na her mespace. I don't think this is unduly
> > overloading it; it's being used as an identifier for the semantics of
> > the tags it describes.
> 
> I think I'd like to know what the XML Namespace people think of the issue. Anyone
> out there?
> 
> >
> >
> > Using a separate identifier doesn't really add anything, unless it's
> > felt that there is an additional, orthoganal way needed to describe
> > the desired handler. It bloats the message, and increases the amount
> > of administrative details associated with messages (a namespace URI
> > and a module-functionality URI).
> 
> Remember we are using XML, so we are not too concerned by the size of messages
> anyway!  :)
> 
> > Do we have any use cases (documented or not) where handlers (not
> > processors) need to be targetted in this manner?
> 
> I'd be tempted to turn the question round: do we have any use case where
> processors (not handlers) should be targetted, considering that handlers will do
> the work anyway? (Am I answering a different question?)
> 
> Cheers,
> 
> Jean-Jacques.

Received on Wednesday, 21 March 2001 10:41:09 UTC