Re: [sca-bindings] [Issue 2] Scope of our charter and ws-callback [was: Proposalversion 3: How to implement SCA callbacks using SOAP]

[Mike Edwards <mike_edwards@uk.ibm.com>]

Folks,

My opinions on this:

1) I believe that it is in scope to
deal with this Issue within the SCA TCs and that Anish's proposal meets
the criteria for being in scope also.
Anish picks out the relevant parts of
the charter below.  I certainly envisaged the use of existing mechanisms
in SCA-specific ways as
being part of what the bindings TCs
would do - hence the specific wording about callbacks and conversations.

2) I believe that this specification
can and should define what is necessary to enable callbacks to work over
Web services.

3) I don't think that this TC should
get engaged in a "WS-Callback" specification.
I can agree that such a specification
might be useful to the wider web services community, but I don't agree
that this TC is the venue
to create and standardize such a specification.


Yours,  Mike.

Strategist - Emerging Technologies, SCA & SDO.
Co Chair OASIS SCA Assembly TC.
IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.
Phone & FAX: +44-1962-818014    Mobile: +44-7802-467431  
Email:  mike_edwards@uk.ibm.com

From:	Anish Karmarkar <Anish.Karmarkar@oracle.com>
To:	David Booz <booz@us.ibm.com>
Cc:	sca-bindings@lists.oasis-open.org
Date:	06/11/2008 07:00
Subject:	Re: [sca-bindings] [Issue 2] Scope of our charter and ws-callback [was: Proposal version 3: How to implement SCA callbacks using SOAP]

---

David Booz wrote:
> Thanks again Anish. I still don't buy your argument about charter
scope 
> WRT normative text, but we can put that aside for a moment.
> 

Starting a different thread (with a different subject) so as to not mix
the technical and scope arguments.

I checked the charter [1]. It does say the following:
"Note that the list of out-of-scope features and mechanisms is not
exhaustive. In general, areas that are not explicitly listed as within
the scope should be considered as out-of-scope."

So, for something like ws-callback to be in scope, it (the 
functionality) has to be explicitly listed in the charter.

In the scope section it says:
"The TC will not explore areas with general applicability and for
which 
typically standards exist."

WS-Callback is of general applicability, but there isn't a standard 
around. There is a spec [2] for it, but it has been superseded by 
WS-Addressing W3C Recommendation. This sentence in itself doesn't put 
ws-callback in scope, but the charter goes on to say:

"For each binding specification, the TC shall identify the applicable
subset and specify solutions accordingly."

and then goes on to list a bunch of things that are in scope. One of 
which is:

"Support for bi-directional and conversational interfaces (as defined
in 
the SCA Assembly specification)"

This arguably puts this in scope. Interpretation may of course differ.
I 
think, if the TC really wants to do this there are ways of clarifying 
the charter.

My opinion is that, if we want to do this, we should do this right: 
create a separate spec that specifies the protocol and the metadata 
(WSDL extensibility, WS-Policy) associated with it. There is nothing 
about it that is SCA specific. I don't see why we would not want this to
interop with other non-SCA systems. For interop (within SCA or outside
of SCA), the spec needs to be well-defined and not a spec by example. 
Decoupling it from the SCA spec sends the right message to the WS 
community and will receive review from the right people. Burying a WS-*
spec (which is really what we are writing, independent of whether it 
appears in some appendix or a standalone spec) in an SCA binding spec 
when it is not SCA-specific will mean that lot of WS-Heads will miss it
and is an under-the-radar way of getting a WS-* spec out. I would 
propose that we should either do it as a separate spec or not do it at
all. I do understand the schedule argument and I don't think we need to
put this spec on the same schedule as the rest of the SCA specs.

My 2 cents.

-Anish
--


[1] http://www.oasis-open.org/committees/sca-bindings/charter.php
[2] 
http://www.w3.org/Submission/2004/SUBM-ws-messagedelivery-20040426/#callback-pattern

> Is the following allowed given your proposal? I think it is, just
want 
> to check. The S messages in your examples look too much like traditional
> responses, something which we need to be sensitive to. And in case
> anyone asks, R is not multi-threaded.
> 
> R invokes YouRIt(). Let's call this invocation R4 and it sets the
> callback address to RC3.
> R invokes YouRIt(). Let's call this invocation R5 and it sets the
> callback address to RC3.
> S then calls NoYouRIt(): S6.
> 
> R4:
> 
> <soap:Envelope ...>
> <soap:Header>
> <wsa:From>
> <wsa:Address>_http://example.com/callback-yetanother_ 
> <http://example.com/callback-other></wsa:Address>
> <wsa:ReferenceProperties>
> <myNS:SomeID>3</myNS:SomeID>
> </wsa:ReferenceProperties>
> </wsa:From>
> <wsa:MessageID>urn:uuid:f81d4fae-10dec-11d0-a765-00a0c91e6bf6</wsa:messageID>
> ...
> </soap:Header>
> <soap:Body>
> ...
> </soap:Body>
> </soap:Envelope>
> 
> R5:
> 
> <soap:Envelope ...>
> <soap:Header>
> <wsa:From>
> <wsa:Address>_http://example.com/callback-yetanother_ 
> <http://example.com/callback-other></wsa:Address>
> <wsa:ReferenceProperties>
> <myNS:SomeID>3</myNS:SomeID>
> </wsa:ReferenceProperties>
> </wsa:From>
> <wsa:MessageID>urn:uuid:f81d4fae-11dec-11d0-a765-00a0c91e6bf6</wsa:messageID>
> ...
> </soap:Header>
> <soap:Body>
> ...
> </soap:Body>
> </soap:Envelope>
> 
> S6:
> 
> <soap:Envelope ...>
> <soap:Header>
> <wsa:To>_http://example.com/callback-yetanother_ 
> <http://example.com/callback-other></wsa:To>
> <myNS:SomeID>3</myNS:SomeID>
> <wsa:RelatesTo 
> RelationshipType=_"http://docs.oasis-open.org/opencsa/sca-bindings/callback"_>
> urn:uuid:f81d4fae-10dec-11d0-a765-00a0c91e6bf6
> </wsa:RelatesTo>
> ...
> </soap:Header>
> <soap:Body>
> ...
> </soap:Body>
> </soap:Envelope>
> 
> 
> 
> Dave Booz
> STSM, BPM and SCA Architecture
> Co-Chair OASIS SCA-Policy TC and SCA-J TC
> "Distributed objects first, then world hunger"
> Poughkeepsie, NY (845)-435-6093 or 8-295-6093
> e-mail:booz@us.ibm.com
> 
> Inactive hide details for Anish Karmarkar ---10/29/2008 07:51:53 
> PM---Attached. I have converted the HTML version to Word, so tAnish
> Karmarkar ---10/29/2008 07:51:53 PM---Attached. I have converted the
> HTML version to Word, so that it is easy for folks
> 
> 
> From:                
> Anish Karmarkar <Anish.Karmarkar@oracle.com>
> 
> To:                
> OASIS Bindings <sca-bindings@lists.oasis-open.org>
> 
> Date:                
> 10/29/2008 07:51 PM
> 
> Subject:              
  
> [sca-bindings] [Issue 2] Proposal version 3: How to implement SCA
> callbacks using SOAP
> 
> ------------------------------------------------------------------------
> 
> 
> 
> Attached.
> I have converted the HTML version to Word, so that it is easy for
folks
> to comment and make inline changes.
> 
> The big change between v2 and v3 is the addition of faults. My initial
> inclination was to define our own faults, but it turns out WS-Addressing
> has some pre-defined faults which fit our purpose. So, I have reused
> those faults.
> 
> One thing I'm not clear on is whether the TC wishes to define this
> SOAP-specific callback mechanism as *one* of the ways callbacks can
be
> implemented when using SOAP or *the* way callbacks are implemented
when
> using SOAP. My inclination is to say *the* way, otherwise we have
to
> invent ways (metadata such as policySets, ws-policy assertion, WSDL
> extension, Java annotation) to convey when it is required.
> 
> Comments?
> 
> -Anish
> --
> /(See attached file: 
> SCA-bindings-issue2-proposal-v3.doc)/---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  Follow this link to all your TCs in OASIS
at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
> 
> 
> ------------------------------------------------------------------------
> 
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  Follow this link to all your TCs in OASIS
at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php

---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php

---

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU