Re: Internet-Draft: Media Feature - xmlns

From: Ian Graham <[email protected]>
Subject: Re: Internet-Draft: Media Feature - xmlns
Date: Thu, 24 Jan 2002 23:45:29 -0500 (EST)

> 
> 
> On 24 Jan 2002, Simon St.Laurent wrote:
> 
> > 
> > On Wed, 2002-01-23 at 22:14, Ian Graham wrote:
> > > My only suggestion, for what it's worth, would be to allow for the
> > > declaration (and please don't overinterpret that word ;-) ) of the 'root'
> > > namespace for a document.  That seems a reasonable distinction to make,
> > > although you'd then have to define how such a statement interacts with
> > > content-type declarations such as application/xml+mathml (would that, in
> > > some twisted way, imply a default 'root'-level namespace, and if so,
> > > how?).
> > 
> > Personally, I believe that the root namespace is not a definitive guide
> > to the contents of the document.  The XSLT case is prominent, but there
> ...
> 
> Well, in practice there is only one namespace distinction you can make --
> between the 'root' namespace and all others.  Specifying both classes thus
> provides, _all_ the namespace information you can provide in the absence
> of the complete document instance.  It seemed to me silly to avoid
> providing this information when it's so easy to (notationally) do so. 

I do not think that importance of a "root" namespace is that 
obvious.

Some people (including me) would like to allow an XML document to have
a sequence of "root" elements rather than a single root element.  Such
XML will be much more appropriate for log files.  I even heard that
one widely-used XML parser already has a hidden option to allow such
XML documents.  If such an extension becomes part of XML, a document
may have more than one top-level namespace.


Cheers,

Makoto