Re: DOM3 Core: Namespaces Algorithms vs xml: and xmlns:

On Dec 19, 2005, at 10:50 AM, Joseph Kesselman wrote:

>
>> would seem quite wrong to be using xml or xmlns prefixes in content.
> I'm not so sure about xml. For example, an XPath might be looking for
> instances of xml:space or similar reserved attributes which do  
> become part
> of the data model. XPath does use qnames to specify such searches, and
> XPaths may appear in content.
>
> XPath doesn't consider namespace declarations to be attributes, so  
> I don't
> have a similar usecase there. I'm not sure that it can't arise,  
> though.
>
> (Comments about the use of implicitly-bound prefixes in content  
> withheld; I
> lost that argument long ago.)

You are right, of course.  Once there is a precedent of representing  
QNames this way in content, who is to say what QName will be  
represented.  I could probably even build a one for xmlns, and there  
is nothing preventing it from occurring in content.

Ray