RE: Proposal for resolving 144, 161, 117: array serialization

 Andrew,
 I agree with Rich's and Noah's replies to the post below.
 I don't see the SOAP Encoding layer as losing any information if
it gives you an array with typed but unnamed values. The names of
the element can imply a type and this information is not lost.
 If we start from a data structure, e.g. an array of typed
unnamed values, and we encode it as a SOAP Encoding array, then
we decode it, adding the name information would actually add
information, which is arguably useless because it was not present
in the original data.
 Let's keep in mind that SOAP Encoding is used for encoding and
then decoding data graphs where the name information is not
present, not for decoding of XML data created in a different way,
in which case preserving the name information could be
meaningful.
 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)
                   http://www.systinet.com/



On Thu, 27 Dec 2001, Andrew Layman wrote:

 > I respectfully disagree that a SOAP layer that loses such information
 > could be described as "compliant" if by that is meant compliant with the
 > full SOAP specification in a general-purpose way.  (The absence of full,
 > general-purpose compliance may be perfectly satisfactory for many
 > applications; that is a different matter.)
 >
 > First, the SOAP specification explicitly sanctions the use of a post
 > schema validation infoset.  See, for instance, SOAP 1.1 section 5 rule 2
 > part c. I believe that the proposed SOAP 1.2 is even more strong and
 > clear on this point.
 >
 > Second, the encoding rules describe mapping between XML instances and
 > graphs.  If a particular SOAP support library happens to lose some
 > information along the way, information that is clearly described as part
 > of the mapping, then the support library does not support that part of
 > the spec, not in a general purpose way.
 >
 > -----Original Message-----
 > From: Rich Salz [mailto:rsalz@zolera.com]
 > Sent: Wednesday, December 26, 2001 5:46 PM
 > To: Andrew Layman; jacek@systinet.com; noah_mendelsohn@us.ibm.com
 > Cc: xml-dist-app@w3.org
 > Subject: RE: Proposal for resolving 144, 161, 117: array serialization
 >
 > >While the element names are not significant qua array element, they may
 > >be significant in other ways, such as affecting the post schema
 > >validation infoset, I would think.
 >
 > I don't think you can count on that, when the SOAP layer assigns its own
 > names or strips them before passing the data up, that's compliant.
 > 	/r$
 >

Received on Friday, 4 January 2002 04:33:51 UTC