Re: [wsdm] WS-Addressing Effects on WSDM

[Mark Little <mark.little@arjuna.com>]

William, like WS-DM, we reference WS-Addressing in WS-CAF and use an 
open content model (because we also support WS-MD as an alternate EPR 
implementation). Now it's always been my understanding from when we 
started WS-CAF that under OASIS rules we would not be allowed to propose 
for adoption a specification that relies on other specifications that 
are themselves not standards in some recognised standards body. At least 
that's what I've been told. Since there isn't a dependency on 
WS-Addressing, we should be able to have WS-Context (as an example from 
the CAF work) pushed to standard level without any conflict.

Now currently we're looking at using WS-BaseFaults from WS-RF within 
WS-CAF. One issue with this would be that we couldn't move for adoption 
of a specification as a standard anything that used WS-BaseFaults until 
that was itself a standard. (Assuming the interpretation of OASIS rules 
is correct; I have no idea how circularities are dealt with!)

WS-MD has a dependency on WS-RF and WS-N. So irrespective of the 
WS-Addressing issue, this still remains. True?

Mark.


Vambenepe, William N wrote:

> A related concern I've heard is that WSDM uses 2 different versions of 
> WS-Addressing. More precisely, WSDM uses one version of WS-A (2004/08) 
> but it also uses WS-BaseNotif 1.2d3 which uses another version of 
> WS-Addressing (2003/02). This is true and of course everyone agrees 
> that it is less than ideal.
>  
> But while this creates a little bit of added complexity for some 
> implementations, in no case does it result in interoperability 
> problems. Each message exchange described by WSDM which uses an EPR 
> (including those inherited from WS-BaseNotif) specifies unambiguously 
> what version of WS-A is used in the EPR. So yes you might have to 
> support more than one version but you always clearly know what version 
> to use when.
>  
> Here is an example:
>  
> Using WSDM, I learn that my manageable resource A depends on a 
> manageable resource B. Through the WSDM relationship mechanism, I get 
> an EPR (version 2004/08) that points to a manageability endpoint for 
> B. Since I care about the health of A, I decide to register for 
> notifications on B so that I'll be alerted when something goes wrong 
> with B that could affect the health of A. To do so, I use the EPR I 
> have for B (again, using WS-A version 2004/08) to send a "subscribe" 
> message to B. The response to this message contains an EPR to the 
> newly created subscription. This EPR uses the 2003/02 version of WS-A. 
> So I now have two EPRs, one (version 2004/08) to the manageability 
> endpoint to B and the other one (version 2003/02) to a subscription. 
> These are used for different interactions and address different 
> endpoints. There is no confusion as to what version of WS-A to use for 
> what EPR.
>  
> And in any case, smart implementations such as the ones in Apache will 
> be, as the saying goes, "liberal in what they consume and conservative 
> in what they produce". From my understanding, the Apache Hermes code 
> will accept any version of WS-addressing.
>  
> Those who have concerns about these problems are warmly invited to 
> join us in the WSDM interop effort.
>  
> Regards,
>  
> William
>
> ------------------------------------------------------------------------
> *From:* Heather Kreger [mailto:kreger@us.ibm.com]
> *Sent:* Monday, February 28, 2005 12:59 PM
> *To:* wsdm@lists.oasis-open.org
> *Subject:* [wsdm] WS-Addressing Effects on WSDM
>
>
> I just thought I'd try to correct any misunderstandings from Martin's 
> post.  In his post
> he makes a set of assertions and I'd just like to clarify one or two. 
>   The assertion that
> the changes to WS-Addressting are disruptive to the WSDM V1.0 
> specifation.
>
> "In
> particular WS-Addressing is currently being worked on and looks like
> the final version when it finally emerges will be significantly
> different from its various antecedent proprietary versions. In
> particular the debates and changes surrounding reference properties
> and parameters will mean the use of different schema types and usage
> patterns. None of these changes will mean that it can't be used by
> these higher level specifications, e.g. WSDM, etc., but they will need
> to be modified. The current Working Draft of the W3C WS-Addressing Working
> Group [2] includes this status section:"
>
> "This is a draft document and may be updated, replaced or obsoleted by 
>  other documents at any time.
> It is inappropriate to cite this document
> as other than work in progress."
>
> Because WSDM treats the entire EPR as opaque. We only use the EPR as a 
> unit.
> We do not suggest how to create the EPR or ever suggest information be 
> extracted from EPRs.
>
> Therefore, WSDM V1.0 is unaffected by changes in the WS-Addressing 
> specification in the W3C.
> WSDM V1.0 does not reference the WS-Addressing Working Draft, it 
> references the submission:
>
> http://www.w3.org/Submission/2004/SUBM-ws-addressing-20040810/. The 
> submission itself is stable.
>
> It is not changing during development of the specification.  It does 
> not have the same text quoted above in the status section.
>
> WSDM V2.0 is already scoped to include making whatever changes are 
> necessary to use the standard
> versions of WS-Addresssing, WS-Notification, and WS-Resource Framework.
>
> Heather Kreger
> STSM, Web Services Lead Architect for SWG Emerging Technologies
> Author of "Java and JMX: Building Manageable Systems"
> kreger@us.ibm.com
> 919-543-3211 (t/l 441)  cell:919-496-9572
>
begin:vcard
fn:Mark Little
n:Little;Mark
email;internet:mark.little@arjuna.com
title:Chief Architect
version:2.1
end:vcard