Ronald Bourret wrote:
>  Note that there is actually a choice here about how to associate URIs
>  and prefixes. Most specs use namespaces declarations. XML-DBMS uses
>  elements at the start of the document to declare namespaces. This
>  simplifies processing and reduces confusion

This is an important distinction. By providing an application-level
namespace declaration mechanism, you avoid the need to carry around a lot of
unnecessary information about in-scope namespaces. I'm glad to see another
real-world precedent for application-level namespace declarations, because I
too think that QNames in values are eminently useful and that they can be
used without relying on xmlns declarations.

>  but also reduces
>  portability, since you can't move map fragments around using something
>  like the DOM and be guaranteed of correct results.

I'm not sure that it would be very reliably portable. Any processing model
that doesn't consider [in-scope namespaces] to be of significance (because
it assumes that XML namespaces are for putting element and attribute names
in a namespace, how novel) will throw away the prefixes and everything will
break.

<snip/> 

>  And no, there is no going back. QNames in attribute values are too
>  massively useful and easy to use.

Let me rephrase the question: Is it too late to require QNames in values to
be resolved with an application-level namespace declaration?

Evan


-----------------------------------------------------------------
The xml-dev list is sponsored by XML.org <http://www.xml.org> , an
initiative of OASIS <http://www.oasis-open.org> 

The list archives are at http://lists.xml.org/archives/xml-dev/

To subscribe or unsubscribe from this list use the subscription
manager: <http://lists.xml.org/ob/adm.pl>