W3C home > Mailing lists > Public > www-tag@w3.org > March 2006

RE: Review of Authoritative Metadata

From: David Orchard <dorchard@bea.com>
Date: Mon, 27 Mar 2006 14:04:00 -0800
Message-ID: <E16EB59B8AEDF445B644617E3C1B3C9C01425E5C@repbex01.amer.bea.com>
To: "Rice, Ed \(ProCurve\)" <ed.rice@hp.com>, <www-tag@w3.org>

I would say that the WSDL is authoritative in regards to what a contract
is, that is the acceptable inputs and possible outputs.  But there are
many times when an individual message and "the WSDL" may get out of
synch.  For example, if the service is versioned and a client has an old
WSDL.  

Now on to examining the content type in each direction.  I would say
that WSDL has 2 directions - in and out.  On the in-side, the client
provided metadata with an encapsulated container must be considered
authoritative.  If the client says it is a SOAP+XOP/MTOM request, that's
what it is.  On the return side, the metadata received still MUST be
considered authoritative.  If the service says it's application/soap+xml
return, then the message is SOAP 1.2 and not SOAP 1.1 (which is
text/xml).  

There's a difference between authoritative metadata for a given message
versus authoritative metadata for a contract.  The provided metadata for
a message always wins wrt to the message.  

Using the XML/XSD example, there are the same versioning potential
problems, where a given XML might be valid under one Schema but invalid
under another Schema.  The given XML will have a namespace name, and
that binding of the namespace in the xml document to a schema (beit rddl
or some other lookup) will determine the self-describing nature of the
xml document.  

Cheers,
Dave


> -----Original Message-----
> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org] On Behalf
Of
> Rice, Ed (ProCurve)
> Sent: Monday, March 27, 2006 12:23 PM
> To: www-tag@w3.org
> Subject: Review of Authoritative Metadata
> 
> 
> Roy/TAG,
> 
> Some thoughts on the latest Authoritative Metadata paper at
> http://www.w3.org/2001/tag/doc/mime-respect-20060307
> 
> 1 - In Section 2 you state "For Web architecture, a design choice has
> been made that metadata received in an encapsulating container MUST be
> considered authoritative.."
> 	Question:  Is this a 'design choice' or a 'determination'?  A
> design choice seems to leave it open to more interpretation than a TAG
> determination..
> 
> 2 - 3.3 external reference metadata is least authoritative.
> 	Question: In deference to most html, doc types I would agree.
> Is the same true however to xml?  Is the WSDL least authoritative in
> regards to a SOAP message?  I believe by definition the WSDL is THE
> authoritative source as to the format of the doc when it comes to web
> services (please correct me if I'm wrong).  Clearly in this instance
the
> WSDL would specify xml, but the element structure/types within the xml
> are also defined in the meta data of the WSDL.  I would also think the
> same applies to an xml/xsd relationship where the xsd is the
> authoritative source regarding the structure of the xml?
> 
> The first item can clearly be dropped into the 'nit' bucket, the 2nd
> item I'd like to hear your thoughts on..
> 
> -Ed
Received on Monday, 27 March 2006 22:06:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:39 GMT