RE: Minutes, 24 Sept 2003 WS Desc WG FTF

Looser interpretation of frag-ids when media type information is not
available doesn't solve the problem, though I haven't looked up the
proposed changes in detail.  Whether media type information is available
or not is an environmental characteristic.  In some circumstances, media
type information might not be available, but in other circumstances, for
the same URI, various media types might be available.  How do I know
that a fragment Id on a namespace identifies an abstract WSDL component
instead of a subresource of the RDDL or whatever returned by
dereferencing the namespace URI?
 
> -----Original Message-----
> From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]
On
> Behalf Of David Orchard
> Sent: Tuesday, September 30, 2003 4:36 PM
> To: www-ws-desc@w3.org
> Subject: RE: Minutes, 24 Sept 2003 WS Desc WG FTF
> 
> >
> >
> > > -----Original Message-----
> > > From: www-ws-desc-request@w3.org
[mailto:www-ws-desc-request@w3.org]
> > On
> > > Behalf Of David Orchard
> > > Sent: Monday, September 29, 2003 5:44 PM
> > > To: www-ws-desc@w3.org
> > > Subject: RE: Minutes, 24 Sept 2003 WS Desc WG FTF
> > >
> > > I'm confused by what happened on this.  My understanding is that
the
> > WSDL
> > > WG made a decision about what it thought was the best approach for
> > > component designators.  It then asked the TAG what the TAG
> > thought of
> > the
> > > WSDL WG's decision.  The TAG has certainly not said that
> > the WSDL WG's
> > > approach is poor.  In fact, the TAG hasn't said anything.  This
has
> > been
> > > annoying to me, but then the TAG is trying to get to Last
> > Call on the
> > web
> > > arch document so the rationale is reasonable.
> >
> > The idea was to see if there was another approach that was not so
> > clearly broken as the approach in our draft, thus allowing us
> > to improve
> > the situation while waiting for the TAG to make progress
> > (which has not
> > been evident.)  After my recent XInclude experience, I think I
> > understand fragment identifiers and media types better than I did
> > before, which just confirms to me how broken our current approach
is.
> > I've been working on fragment identifier syntax and
> > applications for XML
> > for several years now and the result of my experience is to
> > stay as far
> > away from frag-ids as possible.
> >
> 
> I recall that you have made this assertion before, that frag-ids and
media
> types are broken.  I have my own ideas on some of their problems.  But
> could you elaborate on the reasoning that led you to believe that the
> current wsdl approach is broken?  I again refer to 2396bis which tones
> down the language around interpretations of frag-ids when the
media-type
> is not available which I think ameliorates some of the concerns I had
had.
> I obviously request this from 2 perspectives, better understanding of
> WSDLs approach and for possible discussion in the TAG.
> 
> Cheers,
> Dave

Received on Wednesday, 22 October 2003 15:32:48 UTC