Re: Finishing the XML-tagging discussion

[Graham Klyne <GK@xxxxxxxxxxxxxx>]

At 07:16 PM 3/20/00 -0500, Keith Moore wrote:
> > Requesting yet again.  Could we please have details of this horrible
> > problem inflicted by the -xml suffix?
>
> could we please have details of how conneg expressions or HTTP
> Accept headers could be used to say things like
> "I accept XML documents"?

Here is a strawman ...

(a) revise RFC 2295 (currently experimental) so that rather than use its 
own notations for describing resource variants and client features it uses 
conneg expressions.  RFC2295 introduces an "Accept-features:" header that 
is provided by a client requesting a resource.

(b) register a conneg feature tag to indicate XML content;  I would suggest 
something along the lines of (XML-namespace="<URI>"), where <URI> is an XML 
namespace identifier, as that would allow negotiation to specific XML 
applications.  A distinguished value, say "*", would be defined to indicate 
a generic XML handling capability.  An empty string could indicate XML 
handling without recognition of namespaces (i.e. "classic" XML).

(c) use the proposed Content-feature tag for labeling MIME content.

I believe this approach takes far greater account of the way the XML and 
XML applications are being developed by W3C than simply having a parameter 
that says "this is XML".

#g

------------
Graham Klyne
(GK@xxxxxxx)