RE: Proposed Edits to "Framework" spec for header/body distinctio n.

Hi Chris,

> -----Original Message-----
> From: Christopher Ferris [mailto:chris.ferris@sun.com]
> Sent: 04 December 2001 15:00
> To: Williams, Stuart
> Cc: 'noah_mendelsohn@us.ibm.com'; xml-dist-app@w3.org
> Subject: Re: Proposed Edits to "Framework" spec for header/body
> distinctio n.
> 
> 
> Stuart,
> 
> But the processing of SOAP "blocks" is defined/determined
> by the SOAP spec whereas the processing of Body element
> info items is not covered by, prescribed by or even
> hinted at by the SOAP spec except for the case of the
> part2 RPC which provides one way of handling the contents
> of the SOAP Body.

Ok... I think I get it... basically, we expect to find SOAP header blocks in
<env:Header/> and we define a processing model for processing them, although
we offer no prescription for any particular (type of) header block,
certainly no more prescription than we do for the contents of <env:Body>. We
don't expect to find <![CDATA....]> or text or other mixed content in
<env:Header>.

We have no processing model for <env:Body>. For the most part we have always
'imagined' (or at least I have) contained elements as immediate children,
but as you suggest below <env:Body> could contain text or mixed content. In
any case, we have no model in part 1 sections 2/4 that indicates anything
about how the ultimate recipient should address processing the contents of
<env:Body>, its just there and you have to deal with it as a single unit...
whatever it means and whether in some sense you understand it enough to do
anything useful with it.

[Aside I think Noah retained "SOAP header block" possibly for emphasis
rather than the shorter "SOAP block" - but could be a diff reading problem
on my part]

> As for the messaging framework, we really
> have nothing to say about how the contents (if any) of
> the Body are processed. The contents of the SOAP body
> could be a CDATA section. That looks nothing like, nor
> is it processed in any way like a SOAP block.
> 
> <S:Envelope xmlns:S="http://www.w3.org/2001/...">
>    <S:Body>
>      <![CDATA[yaddayaddayadda]>
>    </S:Body>
> </S:Envelope>
> 
> The above is a perfectly valid, if not very useful
> use of a, SOAP message yet there is no element name and
> there is no namespace, hence it isn't a SOAP block.

Incidentally, I think the envelope schema [1] would allows similarly useless
content in <env:Header>.

Personnally, I had rather felt that the assymmetries between headers and
bodies were more related to the way that they were processed than to their
syntactic structure. ie. Both <env:Header> and <env:Body> contain possibly
zero length sequences of blocks, the treatment of those block differs
between header and body, but the syntactic structure (except actor and
mustUnderstand) is/was basically the same.

I still have a preference for a structurally symmetric view even if the
processing is assymetric. 

> Cheers,
> 
> Chris

Regards

Stuart
[1] http://www.w3.org/2001/09/soap-envelope

<snip/>

Received on Tuesday, 4 December 2001 11:47:25 UTC