RE: i037: Replace QName's with anyURI

Let's me illustrate better readability with Harris' example [1]

Look at the EPR as is defined today.
<wsa:EndpointReference xmlns:wsa="..."

Is this more readable than what's above?
<wsa:EndpointReference xmlns:wsa="...">

When you look at the PortType, you can see both the port type and the
namespace, it's right there. In the current you need to scan the xml
document to look for xmlns:fabrikam and what then figure out what
namespace that it is mapped to....



-----Original Message-----
[] On Behalf Of David Orchard
Sent: Friday, December 03, 2004 2:58 PM
To: Hugo Haas;
Subject: RE: i037: Replace QName's with anyURI

There's been no proof that the problems that the TAG finding talks about
are relevant to WS-A.  I generally refuse to do things "because somebody
smart said so", and *especially* when the smart people didn't outlaw
QNames and noted many cases where they were useful.  I remember when I
mentioned to TimBL that OASIS WS-Security had moved to URIs from QNames,
and he sighed and said "there goes readability".


> -----Original Message-----
> From:
>] On Behalf Of Hugo Haas
> Sent: Friday, December 03, 2004 11:35 AM
> To:
> Subject: Re: i037: Replace QName's with anyURI
> * David Orchard <> [2004-12-03 06:56-0800]
> > In general, +1.  It seems to me that any rationale for moving part
> > WSA QNames to URIs would be to provide some kind of benefit.  I'm
> > strongly against moving relationshipType to URIs, but I'd like a 
> > stronger reason than "because".
> I think that the motivation is not just "because", but the TAG finding

> on QNames that I'm sure you're familiar with:
> | In so far as the identification mechanism of the Web is the URI and 
> | QNames are not URIs, it is a mistake to use a QName for
> | when a URI would serve.
> As this is internal to Addressing, it seems like a simple and natural 
> change to do.
> So I quite like Harris's proposal.
> Regards,
> Hugo
> --
> Hugo Haas - W3C
> -

Received on Friday, 3 December 2004 21:11:48 UTC