Arjun Ray wrote:
>  Paul Prescod <paul@[...].net> wrote:
>  | Arjun Ray wrote:
>  
>  |> "It can't be done" is an acceptable answer, btw.
>  | 
>  | There is no global rule about the semantics of embedded namespaces. 
>  
>  Semantics ("what names mean") wasn't the issue.  If you mean that there is
>  no global rule about the generic semantics of embedding syntax (using the
>  namespaces device) then I agree.  That was, after all, my question. :-)

Right. You asked a question and I provided an answer.

>  | Any rule that handled the book/html case would fall apart in the face of 
>  | XSLT/XSLFO.
>  
>  The general problem is the same.

Not really. One is a code/literal problem. The other is a "reuse these 
semantics" problem. They are quite different which is why there is no 
general semantics for embedding namespaces.

>  | Nevertheless, they have found widespread use as a trigger for 
>  | specific processors, like XSLT engines.
>  
>  Actually, they have found widespread use for DWIMming. 

It isn't DWIMming. XSLT is very precise about how to interpret 
"extension elements" versus "literal result elements" versus "built-in 
elements". WSDL is also very precise. RDF is also very precise. They 
just don't all use the same rules. They ascribe their own semantics to 
the syntactic convention of "foreign namespaces".

>  ... The asymmetry in
>  John's first stab at a procedure is an example.   The real concern there
>  was "knowing" when to leave out #PCDATA.

If BookML were formally defined then the formal definition should tell 
you how to handle PCDATA in the extraction of the implicit XHTML 
document. People are even working on languages specific to that 
extraction problem. But the simple answer is you cannot tell how to 
segment namespaced fragments just by relying on the semantics defined in 
the Namespaces specification. You must also rely on the rule defined by 
your vocabulary.

Actually, it gets worse. If you are presented with an XML document and 
no external type metadata, there is no way to know which namespace to 
use as the one that defines the fragmentation semantics.

  Paul Prescod


-----------------------------------------------------------------
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>