- From: Doug Davis <dug@us.ibm.com>
- Date: Tue, 29 Jan 2002 08:40:07 -0500
- To: "Jean-Jacques Moreau" <moreau@crf.canon.fr>
- Cc: "Noah Mendelsohn" <noah_mendelsohn@us.ibm.com>, "'Henrik Frystyk Nielsen'" <henrikn@microsoft.com>, skw@hplb.hpl.hp.com, xml-dist-app <xml-dist-app@w3c.org>
If the bindings are not held to same processing rules/restrictions (and are separate from the SOAP node) then it seems perfectly valid for them to modify the envelope (e.g. add headers...) just as long as they are removed before the envelope is handed over to the next node. That adheres to the "main rule" Noah has put forward - infoset remains unchanged - and it seems less complicated than nesting envelopes - this leaves it open for bindings to use the envelope (or not) - just as long as they adhere to the "main rule". -Dug "Jean-Jacques Moreau" <moreau@crf.canon.fr>@w3.org on 01/29/2002 08:19:51 AM Sent by: xml-dist-app-request@w3.org To: Noah Mendelsohn/Cambridge/IBM@Lotus cc: Doug Davis/Raleigh/IBM@IBMUS, "'Henrik Frystyk Nielsen'" <henrikn@microsoft.com>, skw@hplb.hpl.hp.com, xml-dist-app <xml-dist-app@w3c.org> Subject: Re: Resolving the Ed Note in Part 1 section 5.1 (was New Issues) Presumably, one could use nested SOAP envelopes to get around the problem of not being able to apply the SOAP extensibility framework. In this model, the initial envelope would be wrapped into a second envelope that would be delivered to the next hop. The second envelope would contain binding specific information, represented as headers (bodies?). :) Jean-Jacques. noah_mendelsohn@us.ibm.com wrote: > You raise a good point. In this proposal, the binding is indeed viewed as > separate in the sense that the processing rules of chapter 2 apply >after < > a binding has done the job of receiving an infoset, and at an intermediary > >before< the relayed infoset is sent by the binding. So, in that sense > separate. > > The proposal I made is intended as a compromise. By imposing the > separation, we get out of the business of figuring out how to integrate > the two. For example, we don't have to say how a binding can munge with > the envelope when in fact the processing rules say that >all< mU checking > must be done before any processing is done. What we lose is the ability > to apply the soap extensibility and processing model to bindings. > > ------------------------------------------------------------------ > Noah Mendelsohn Voice: 1-617-693-4036 > IBM Corporation Fax: 1-617-693-8676 > One Rogers Street > Cambridge, MA 02142 > ------------------------------------------------------------------
Received on Tuesday, 29 January 2002 08:40:16 UTC