[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded
Igor, Taking a step back, I believe that using the singleton pattern is compliant with WSRP. I also believe it is OK for the April 14 interop to use only the singleton pattern. It is a good simplifying choice that allows us to get something interoperating more quickly. I also think that omitting the WS-Addressing SOAP headers is another simplification that allows us to get something working more quickly and is a good choice for our interop. I am less thrilled with allowing wsrp:ResourceProperties to be omitted from the WSDL. Tools that consume WSDL will ignore it if they don't understand it so there is no reason not to have it present for them. Tools that generate WSDL are not as important for our current interop and will need to be versioned in order to support WS-RP. However, I am OK with this decision by the group for the purposes of achieving interoperability. Regarding WSRP compliance: Using the singleton pattern does not mean that you are allowed to stop using the WS-Addressing 'wsa:To' and 'wsa:Action' headers. WS-Addressing requires these 2 headers. A Web service compliant with WS-RP must be compliant with WS-Addressing: "A message that follows the implied resource pattern MUST be sent to a Web service referred to by an EndpointReference that follows the implied resource pattern, and MUST conform to the WS-Addressing requirements on that message including adding ReferenceProperties information, if present, from that EndpointReference to the message." (WS-RP version 1.1 3/5/2004, section 2, Implied Resource Pattern, bullet 3) WS-Addressing requires that the wsa:To and wsa:Action headers be included in messages. Using a singleton pattern does mean that request messages will not contain any SOAP header that is used for selecting a WS-Resource. I don't understand how a WSDL transform has anything to do with complying to WSRP. You either expose compliant WSDL or you don't. It is not up to a WSDL consumer to transform it before reading it. Furthermore, I don't believe that it is possible for a transform to decide which property document to associate with which portType. In summary: I think the WSDM 0.5 specifications need to be compliant with WS-RP and associated specs and believe that they are now. I also think that we need to make some simplifying assumptions in order to get started on interoperability and I believe that we have done that. Bryan -----Original Message----- From: Sedukhin, Igor S [mailto:Igor.Sedukhin@ca.com] Sent: Tuesday, March 30, 2004 5:57 PM To: Murray, Bryan P.; Vambenepe, William N; wsdm@lists.oasis-open.org Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded 1) well if one can take that WSDL and autotransform it to the one that WSRP requires, that IMO, makes the overall case compliant. The WSDL reduction is merely our own technical simplification for the interop. 2) And while I was not able to identify precisely if WSA header is required in WSRP messages or not from (http://www-106.ibm.com/developerworks/library/ws-resource/ws-resourcepr operties.pdf). The following statement [ A stateful resource identifier MAY be contained within the ReferenceProperties element of the Endpoint Reference. ] This means that it may also not be contained there, which is the case for the April interop. -- Igor Sedukhin .. (igor.sedukhin@ca.com) -- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788 -----Original Message----- From: Murray, Bryan P. [mailto:bryan.murray@hp.com] Sent: Tuesday, March 30, 2004 6:02 PM To: Sedukhin, Igor S; Vambenepe, William N; wsdm@lists.oasis-open.org Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded Regarding April 14 interop test compliance with WS-RP: 1) WSDM interop says the ResourceProperties attributes may be omitted - this is how a property document is associated with a portType in WS-RP. 2) None of the example messages for WSDM interop include WS-Addressing SOAP headers and this morning on the interop call we decide that for April 14 they would be optional and could not include the mustUnderstand attribute set to true. WS-RP requires an implementation to comply with WS-Addressing (including its use of SOAP headers). Bryan -----Original Message----- From: Sedukhin, Igor S [mailto:Igor.Sedukhin@ca.com] Sent: Tuesday, March 30, 2004 2:35 PM To: Murray, Bryan P.; Vambenepe, William N; wsdm@lists.oasis-open.org Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded Then we must delay publication of WSDM 0.5 until it clarifies how exactly to get proper EPRs. Otherwise it is unuseable. The statement in the WSRF [ A WS-Resource-qualified endpoint reference may be returned as a result of a Web service message request to a factory to create a new WS-Resource or, alternatively, from the evaluation of a search query on a service registry, or as a result of some application-specific Web service request. ] Does not help WSDM 0.5 implementers at all. The http://www-106.ibm.com/developerworks/library/ws-resource/ws-modelingres ources.pdf says [ Note that other patterns for enabling access to stateful resources are possible. For example, a Web service could maintain the resource identity as static service state, thus obviating the need to pass that identity in the WS-Addressing endpoint reference. This design choice implies a one-to-one mapping from Web service endpoints to stateful resources and thus a need for a unique Web service endpoint for each stateful resource. ] So I don't understand why April 14-15th tests are not compliant to WSDM 0.5 or WSRP or WSRF?... It was just a matter of clearly stating the case istead of delaying it for the better times... -- Igor Sedukhin .. (igor.sedukhin@ca.com) -- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788 -----Original Message----- From: Murray, Bryan P. [mailto:bryan.murray@hp.com] Sent: Tuesday, March 30, 2004 5:03 PM To: Sedukhin, Igor S; Vambenepe, William N; wsdm@lists.oasis-open.org Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded WSDM has decided to use WS-ResourceProperties to support its need for attributes for version 0.5. I think we should not add a constraint to the WSDM specs contradicting some of what is required by WS-RP. This does not mean that a manageability endpoint needs to supply an EndpointReference that makes use of ReferenceProperties. But, since WS-RP is based on WS-Addressing, it does mean that a compliant manageability endpoint does need to understand the WSA SOAP headers. Furthermore, a compliant manageability client does need to understand the ReferenceProperties aspect of an EndpointReference and treat them appropriately when sending messages to that endpoint. The interop tests we run in April 14-15 are not WSDM 0.5 compliant. I expect that the tests we run at the next face-to-face will be. We can't just kludge the spec to make our interop tests work. The interop tests need to be used to validate that we are using the right technology in our spec. Bryan -----Original Message----- From: Sedukhin, Igor S [mailto:Igor.Sedukhin@ca.com] Sent: Tuesday, March 30, 2004 1:46 PM To: Vambenepe, William N; wsdm@lists.oasis-open.org Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded I just want to define consistently interoperable specs. This has nothing to do with the interop scenario. If I give you a WSDL that follows WSDM 0.5 and in the binding it states that WSA headers are required and mustUnderstand. There is no way you can talk to such manageabilty endpoint without knowing where to get the EPRs. In the 1.0 or 0.8 if we decide to explain how and where to get the EPRs for this case, we can remove the statement. The problem here is that WS-Resource pattern is the "default" in WSRP and so we'd have to constrain to make it work for 0.5. If it was otherwise, i.e. "singleton" was defualt, we'd not have to state anything. -- Igor Sedukhin .. (igor.sedukhin@ca.com) -- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788 -----Original Message----- From: Vambenepe, William N [mailto:vbp@hp.com] Sent: Monday, March 29, 2004 9:31 PM To: Sedukhin, Igor S; wsdm@lists.oasis-open.org Subject: RE: [wsdm] Groups - wd-wsdm-muws-0.5-20040329-with-tracking.zip uploaded Hi Igor, > 770 Section 6 should say > "WS-ResourceProperties MUST be used in a singleton pattern. WSA > headers are not expected and not required in message exchanges." We > may remove this statement after we have discussed this and addressed > properly. How about we discuss before adding the statement? :-) I don't understand why we need this statement. I understand that we have chosen to limit ourselves to this case in the interop demo, but why does that mean the spec needs to be limited to this? This is just one of many decisions we have made for the sake of simplicity in the interop and I don't have a problem with it in that context. But what is the rational for restricting the WSRF implied resource pattern in the spec? Do you have a reason that would not apply to WSRF but would apply to WSDM and would therefore require us to profile WSRF in such a way? William To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsdm/members/leave_workgrou p.php. To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsdm/members/leave_workgrou p.php. To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsdm/members/leave_workgrou p.php.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]