Re: text/xml for SOAP (and XP) considered harmful

I don't disagree.  My solution is not intended to be a long term
solution, because of the issues you (and TimBL) raise around unnecessary
central coordination.  Once we can reasonably assume the other end of
the pipe can dispatch on XML namespaces, then the media type can forever
be hardcoded as "application/xml". 8-)

MB

Andrew Layman wrote:
> 
> Right. The main constrast is between relying on a special registry of
> subtypes and subpurposes of XML versus having a single, integrated system of
> identifiers, namely URI.  What exactly the URI references is possible then
> to work out; it could certainly be as expressive as anything that avoids
> URIs.
> 
> See "Web Architecture from 50,000 feet" at
> http://www.w3.org/DesignIssues/Architecture .
> 
> -----Original Message-----
> From: Frystyk
> Sent: Tuesday, December 19, 2000 11:20 AM
> To: Mark Baker; mmurata@trl.ibm.co.jp
> Cc: xml-dist-app@w3.org
> Subject: RE: text/xml for SOAP (and XP) considered harmful
> 
> Whether it is easier or not is always a difficult discussion but I agree
> that
> it is a useful mechanism. All I am saying is that I don't want to have to
> register and maintain a token in a central registry and have to provide a
> special parser case in order to do that.
> 
> Henrik
> 
> > I don't believe it only works for single namespace documents, though
> > perhaps for a different definition of "works" 8-).  Being able to
> > MIME-dispatch to a SOAP/XP processor by using the
> > application/[xp|soap]+xml media type means never having to assume
> > dispatching on XML namespaces in the XML processor.
> >
> > Assigning responsibility for XML namespace dispatching to an
> > application
> > processor (XP/SOAP in this case) is a lot easier to deploy than
> > assigning it to XML processors.

Received on Tuesday, 19 December 2000 15:56:36 UTC