Re: Starting the ietf-xml-mime mailing list

["Martin J. Duerst" <duerst@xxxxxx>]

At 15:31 99/04/07 +0900, MURATA Makoto wrote:
> Martin J. Duerst wrote:

> > Define one parameter, maybe named "namespaces", that lists, comma-separated,
> > all the namespaces used in the document. The DOCTYPE, if present, counts as
> > one. The DOCTYPE, or the namespace of the root element, is listed first.
> > If other description mechanisms (schemas,...) are defined, the URIs used
> > for them can also be used.
> 
> I am wondering if it is really a good idea to put every information in 
> the MIME header.

Definitely not.


> How can a MIME processor or vCard application extract a vCard fragment 
> from the XML document?  To do so, somebody has to parse the document.  But 
> if somebody has to parse it anyway, it should be really easy to find 
> which namespace appears in the document.  Thus, there is not much point 
> in having namespace information in the MIME header.

If we apply this argument fully, then the doctype is also not necessary.

I think that, in order to decide which application(s) handle the document,
what we have to think about is things such as:

- What namespaces are included in the document.
- Which of these namespaces have to be understood completely in order
  to process the document reasonably.
- Which of these namespaces can be processed partially, or just ignored.

There may be applications that can e.g. only process pure XHTML, or
XHTML with some other namespaces, or XHTML with any additional namespaces,
and so on.


Regards,   Martin.


#-#-#  Martin J. Du"rst, World Wide Web Consortium
#-#-#  mailto:duerst@xxxxxx   http://www.w3.org