W3C home > Mailing lists > Public > xml-dist-app@w3.org > June 2001

Re: Infoset based rewrite of SOAP Section 4

From: christopher ferris <chris.ferris@east.sun.com>
Date: Sat, 30 Jun 2001 07:07:21 -0400
Message-ID: <3B3DB2E9.81FEFD3A@east.sun.com>
To: Noah_Mendelsohn@lotus.com
CC: Mark Nottingham <mnot@mnot.net>, xml-dist-app@w3.org
Noah,

Very interesting discussion. Some thoughts.



Noah_Mendelsohn@lotus.com wrote:
<snip/>
> A few other comments:  I think we need to follow through on some other
> implications for this to be coherent.  Specifically, because this
> addresses only the abstract Envelope section, a naive reader might wonder
> where the <...> went.  After all, most SOAP messages we've seen actually
> do have <..>.

Indeed. Of course, we can always provide example serializations to XML.

> 
> I think many of us agree that the answer will be found when we describe
> the way bindings work:  any SOAP envelope Infoset necessarily has a
> representation as serialized, well formed XML.  So, I expect that when the
> dust settles, the responsibility of a transport binding will be (roughly)
> to move the envelope Infoset to the next point in the message path, and to
> reconstruct it there.  The binding corresponding to the current SOAP v 1.1
> HTTP binding will indeed use the standard serialization in well formed
> XML: it will have the angle brackets that everyone expects.  Indeed, its
> messages may be nearly indistinguishable from those used today for SOAP
> v1.1.  Lots of other bindings will make the same choice, but others such
> as those that do compression will have wire formats that use other forms
> for the Infoset.

This is where things get a little tricky IMO. How is the serialization 
mechanism conveyed from the sender to receiver. How is it agreed upon?
Should the XML declaration <?xml version="1.0"?> be required on all
messages? Is the
MIME Content-Type header adequate to represent the infoset serialization
mechanism applied? 
 
Again, the idea is compelling in some regards, but it needs some further
thought.

> 
> I think that those of us who are enthusiastic about the Infoset
> formulation roughly agree on the above.  I am only pointing out that,
> without the discussion of bindings, readers may be confused.  In summary,
> I really like where this is going, but I don't think it quite stands alone
> yet.  Gudge:  thanks so much for moving this ahead.

+1

> 
> ------------------------------------------------------------------------
> Noah Mendelsohn                                    Voice: 1-617-693-4036
> Lotus Development Corp.                            Fax: 1-617-693-8676
> One Rogers Street
> Cambridge, MA 02142
> ------------------------------------------------------------------------
Received on Saturday, 30 June 2001 07:13:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:01 GMT