Re: Fwd: Proposed LC review of SOAP Attachment Feature doc

Brian,

I'm not sure what we'd say about this.  I looked at the working draft a 
while back and while it seemed OK as far as it goes, I didn't notice any 
special relevance to RDF.  Here are the comments I made to a different 
(non-W3C) forum.

[[
In the medium/long term, I think it may have significant implications for 
the transport of XML data, as it starts to address one of the fundamental 
limitations of XML as a protocol framework, viz the difficulty/awkwardness 
of including non-textual data in an XML envelope structure.  SOAP is one 
such envelope structure, but there are others proposed.

Note that some XML-based protocol frameworks (e.g. BEEP, APEX) address this 
problem by explicitly by embracing MIME structures to enclose the XML 
content.  But many others (notably SOAP) don't.

I think the framework described by the SOAP attachment feature is OK as far 
as it goes, but it explicitly does not specify a particular structure for 
compound messages, other than a vague hand-wave in the direction of MIME 
and DIME.  One of the concerns noted is to avoid interference between the 
MIME encapsulation and the XML content.  I think it can be argued that many 
features have been incorporated into MIME that may have been reasonable in 
MIME's original role but which are not really needed for a general compound 
message format.

If one adopts the view that MIME is going to be a significant application 
protocol element for the foreseeable future, I think there may be a case 
for constructing a profile of MIME that focuses only on those features 
relevant to message transfer, and dropping the "metadata" and other 
structured information that sits more comfortably in an XML envelope 
structure.  Thus, we could end up with a layering something like this:

   +----------------------+
   | application data     |
   +----------------------+
   | XML message metadata |
   +----------------------+
   | MIME encapsulation   |
   +----------------------+
   |     :                |
   |   (wire protocols)   |
   |     :                |
   +----------------------+

(Which is pretty much what one would end up with by transferring SOAP over 
HTTP.)
]]

#g
--

At 10:29 AM 9/20/02 +0100, Brian McBride wrote:

>Looks like another last call doc or review from the XMLP folks.  Any 
>volunteers to review on behalf of the WG?
>
>Brian
>[...]
>
>>The XML Protocol WG is planning to submit one document, "SOAP 1.2
>>Attachment Feature", for Last Call review. The abstract reads:
>>
>>This document defines a SOAP feature that represents an abstract model for
>>SOAP attachments. It provides the basis for the creation of SOAP bindings
>>that transmit such attachments along with a SOAP envelope, and provides for
>>reference of those attachments from the envelope. SOAP attachments are
>>described using the notion of a compound document structure consisting of a
>>primary SOAP message part and zero or more related documents parts known as
>>attachments.
>>
>>The document is short, about 8 pages including references and boilerplate,
>>and does not specify any particular implementation for attachments. The LC
>>document will be a very slightly modified version of the current WD [1]. We
>>would especially like reviews from the following WGs: I18N, QA, Semantic
>>Web (WG's therein), Web Services Architecture, Web Services Description,
>>XForms, XML Core, XML Encryption, and XML Signature. My question to those
>>WGs (and to any other interested WGs) is whether you would be _unable_ to
>>provide a review if the LC period started around Sept 23 and ended around
>>Oct 14.
>>
>>Thank you,
>>David Fallside
>>XMLP WG Chair
>>
>>
>>[1] http://www.w3.org/TR/2002/WD-soap12-af-20020814/
>>............................................
>>David C. Fallside, IBM
>>Ext Ph: 530.477.7169
>>Int  Ph: 544.9665
>>fallside@us.ibm.com
>
>-------------------
>Graham Klyne
><GK@NineByNine.org>

Received on Friday, 20 September 2002 07:31:17 UTC