Web Services Resource Access Working Group Teleconference -- 27 Jan 2010

<trackbot> Date: 27 January 2010

<Bob> trackbot, start telecon

<trackbot> Meeting: Web Services Resource Access Working Group Teleconference

<trackbot> Date: 27 January 2010

<Tom_Rutt> what is the conference code today?

<Bob> 9772

<Bob> scribenick: Sreed

Bob: Minutes of meeting - Jan 19 is approved
... WSRA F2F is coming up with Redmond Microsoft campus Mar 30th to April 1st
... Any objection related to WSRA F2F meeting to be scheduled on Mar 30th to Apr 1st
... Follow-up F2F end of May

Ram: Last week of Jun is fine (22nd to 24th)

Yves: will host next F2F meeting

<dug> s/will/can/ :-)

Bob: The next F2F proposed in Jun 15th to 17th
... F2F meeting - Jun 15th to 17th in Sophia, France
... http://www.w3.org/Bugs/Public/show_bug.cgi?id=8205

8205

dug: Minor spec unique more than one meta resource is associated - MEX Data element

Asir: Need to understand the value

dug: Two meta data resources both fo them MEX meta data element - two docs, not able to combine need to keep it separate

Asir: MEX data container attribute extensibility that changes the semantics?

dug: Adding an attribute extensibility addtionally qualifies in the MEX data element - information

Asir: Need to look into a concreate example on this

Ashok: what would be situation related to this

dug: merge meta data elements together

Asir: I am not seeing much value to this
... I would like to undertand the value

DaveS: What semantic would you attach to the spec in related to this

Asir: If you are changing the semantics if you are adding/changing - relationship - just modfied the representation

dug: MEX meta data create a sub class - extensibility point changes to semantics of sub class
... not changing the basic semantics

Gil: logical conclusion - section level it should be fine

dug: there is inconsistent in the spec now

jeffM: addiing semantics doesnt change the scope

dug: Trying to understand what the problem in making the changes

Asir: It can be represented using the exisiting for making the change provide the information that would help to articulate others

Bob: Extensibiility elements - there is no way of scoping the extensions - in this case there is no change in the semabtics of the parent

Gil: merge extended meta data at the section level

Ram: Attribute extensibilty element violates unique constraint then need to discuss

dug: can't merge them in this case

Gil: extend meta data element - fix one of them remove attribute extensibility or unique constraint

Asir: 7986 attaching the policy what you do for attribute extensibility?

Bob: members accepting Dug's proposal

Asir: what is the compelling argument -related to attribute extensbility

Ashok: Technical arugument - lot better in this semantic defintion more than one data section - provide a real world use case example related to this

Asir: I want to have logical reason for the change

Wu: Implementation perspective I don't see much conflict

<Bob> preferred way is to use the multiple mex element approach.

Bob: Issue no 8205 resolved as proposed with the ammendedment - preferred way is to use the multiple mex element approach

<asir> when there are multiple mex:Metadata elements are involved

<Wu> In addition to doug's proposal, the added text can be "It may also embed as multiple metadata separate metadata sections under one metadata element"

<Wu> "It may also embed multiple metadata as separate metadata sections under one metadata element"

8290

<Bob> RESOLUTION: 8205 resolved as proposed plus additional recommendation text

http://www.w3.org/Bugs/Public/show_bug.cgi?id=8290

<Ram> Proposal: http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0161.html

<dug> to the end of bullet #2: , or an HTTP reachable resource.

Bob: Resolved issue 8290

8292

http://www.w3.org/Bugs/Public/show_bug.cgi?id=8292

<Ram> Proposal: http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0042.html

Bob: Any discussion

<Ram> Proposal: http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0092.html

Ram: accepting the proposal retaining the sentense that was removed

<Bob> resolved as amended

Bob: Resolved 8292

8294

http://www.w3.org/Bugs/Public/show_bug.cgi?id=8294

<Ram> Ram's comment: http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0150.html

Ram: comment above

Bob: Any objection
... Issue 8294 resolved with Ram comment

8180, 8299, 8302

<gpilz> http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0162.html

Bob: any discussion on combined proposal as posted by Gil?

gpliz: Resource Manager/Server - intro section discusses related to operations - 3 use cases - added a new paragraph, info as in the above doc

Ram: Send me a request to create a resource & update a request, is there are any expectation - clarification question

gpliz: Next change adding the representation element - nothing in it

Ram: what are the differenent representations

gpliz: As discussed in the doc

Bob: Any discussions

Ram: Discussed in the message - what is null really mean. 3 Cases have been disucssed
... Non-empty representation, Empty-representation matches with gpliz proposal
... Null representation assuming resources does not exsist - how the behavior effected

gpliz: In the proposal it address the create - null or non existent
... Resource associated to null constructor
... Default constructor creating resource with no representation - return empty

Ram: reffering to doc - last column Null representation assuming resource exists default form (assumption)
... How do we express the resource which is in default form in XML

Bob: XML resource has representation

<DaveS> http://www.w3.org/TR/REC-xml/#sec-well-formed

Bob: Gill's proposal few areas need to be modified or clarified

<Bob> we resume

6435

http://www.w3.org/Bugs/Public/show_bug.cgi?id=6435

gpliz: Send renew request, expiration time goes off - pending expiration - end renew response, service the description (active)

Bob: State table of Subscription/Event state table
... Subscription expires during the renew - inconsistent state of the subscription - renew occurs

gpilz: this is distributed computing all the entire subscription/event will have several states

Bob: State to change is timeout other is renew, if renew happens it cant timeout - state of the object

li: I agree with gpilz point - state of the subscription (data structure represents the subscription)

dug: two diff subscription objects but they are separate objects - going with Li direction these dont have the coorelation services v/s client

Bob: Li can discuss on your second point
... we have series of operations - respone might be message or state change,but one side is the state of subscription , it is not state sharing protocol - stable protocol which could specify known state

Li: very good suggestion

<li> http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0173.html

Bob: Pointng to event source subscription table - subscribe request if not active create, for a diff response to state
... dug's observation subscription before exisits doesnt exist, is the state of non-existent - no distinction one time existed

dug: state table about an event source - one row - subscription state table doesnt have subscribe request row

Bob: Issue getting lifecycle - subscription manager job

<Wu> There is no distinction in terms of the state that a non-existent subscription, and a non-existent subscription but one time existed (e.g. before expired)

<Tom_Rutt> can anyone clarify why it is bad to separate the state tables, one for each processing entity subject to conformance, Why combine two into one table?

<Tom_Rutt> answer, It is an editorial convenience, since one of the tables is so trivial

Tom Rutt: will it be a maintenance issue related to this proposal

<Tom_Rutt> Is the benefit of the state tables being present to help the reader, worth the effort to maintain the spec over time as changes are made to the text

Ram: fine with the direction the challenge - from implementation perspective the developers might feel different

Wu: I second ram's comments & also second Bob's comment, this is part of the spec & we need to maintain

<MartinC> +1 to bob

Bob: Continue working on state tables - to be completed

<Wu> State table is informative and we can decide at LC, since it does not add any additional semantics.

Bob: we should continue on disucssions
... Choices - we can accept the draft or required changes to it or defer for now & discuss it later
... related to state table is anything wrong
... related to proposal will be deffered for discussion till tomorrow & issues can open on this if any

<li> bye

7986

<dug> http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/att-0170/wseventing-7986.html

dug: policy alternative - superset messages

gpilz: Subscription - notification WSDL & other set of subscribers provide different choice - is this possible in the proposal

<asir> i have another point to make as well

Asir: set of policies - a policy has no of alternatives

Wu: I agree with Asir & gpilz comments

Bob: Last sentence removed in the proposal any objects

<dug> When present, this OPTIONAL parameter includes the Endpoint Subject Policy that the event source supports for sending notifications. This element MUST have one child element - typically a wsp:Policy element or a wsp:PolicyReference element. A subscriber can use this information to discover the valid set of alternatives that MAY be used within a wse:NotifyTo EPR which will be used for any...

<dug> ...Notification message sent from the event source to the event sink. If the event source advertises Notification WSDL then any Policy associated with those WSDLs SHOULD be consistent with the alternatives in this parameter.

asir: what it means by consistent
... how do you enforce the policy & can drop the last sentence

<Wu> How about "Any policy expressions included within this parameter SHOULD be consistent with those policy expression alternatives s of the Notification WSDL that the event source advertise."

<dug> When present, this OPTIONAL parameter includes the Endpoint Subject Policy that the event source supports for sending notifications. This element MUST have one child element - typically a wsp:Policy element or a wsp:PolicyReference element. A subscriber can use this information to discover the valid set of Policy Alternatives that MAY be used within a wse:NotifyTo EPR which will be used for...

<dug> ...any Notification message sent from the event source to the event sink. Any Policy Alternatives included within this parameter SHOULD be compatible with those Policy Alternatives available of the Notification WSDLs that the event source advertises.

Bob: we should be able to fix this
... Any objection to the new text ammended to the proposal
... Any objection

RESOLUTION: 7986 as described above

7791

<Bob> proposal at http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0180.html

asir: optional policy assertion parameter defined as mentioned in proposal

dug: policy is one many possibe things future client talking to future resource - it should resource metadata minimum include WSDL what operations & attached - abstract them more to metadata

asir: I need to know the policy - transfer resource perspective

dug: limited view current use case care about - what about the specific metadata

DaveS: if the factory is resource is the MEX metadata associated - in the factory metedata find WSDL it comes back - deployment strategy

asir: talking to resource factory - will it give metadata giving about the children
... as discussed yesterday related to event source policy assertion WSDL of the event source for all the notification messages WSDL can carry policy & the metadata

Bob: Any objection to proposal

dug: Doesnt accept the current proposal

<dug> if we adopt this direction then we need to extend it

<scribe> ACTION: Bob to responed to issue submitter [recorded in http://www.w3.org/2010/01/27-ws-ra-minutes.html#action01]

<trackbot> Created ACTION-143 - Responed to issue submitter [on Bob Natale - due 2010-02-04].

<dug> http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/att-0166/wsfrag-8191.html

8191

<dug> http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/att-0166/wsfrag-8191.html

<dug> In the event that there is more than one node which would match the XPath, the implementation MUST select or return the first node only.

<Bob> ack 8191

<Tom_Rutt> I cannot hear ram now

Ashok: Suppose you are doing put - XPATH selects 5 nodes

Ram: providing expression language XPath 1.0 - check to see we use Xpath on using multiple nodes
... we are taking away some flexibility in XPath

Bob: Need some more time on this?

8196

8229

8196

Tom_Rutt: I dont understand about this proposal - prefix mapping anybody using XPath these prefix are namespaces, I like the way done in CMDBf

gpilz: XPath 1.0 specify local name possible call out text value

dug: explain related to late binding

Ram: no transformation been produced been consumed, there is transformed in the middle transforms or needs to do

asir: I am not fully convinced with this problem talking about XPath there are no. of usage which are succesful XML Query, XML Schema - take care of the mapping of prefix. We have good support SOAP 1.2 it clearly identifies elements must statement be preserved

gpilz: Other specs are broken that needs out spec is broken

<Tom_Rutt> the problem is this: the filter writer wants to write constraints which require the use of namespaces,

<Tom_Rutt> the filter writer does not know what prefixes are going to be used in what is being filtered on

<gpilz> here's a link to a (slightly incorrect) example of using namespace-uri(): http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0130.html

<Tom_Rutt> xpath was designed to work with both xquery and xslt, each of these has a different way to resolve namespace prefix mappings

<Tom_Rutt> so xpath has no way to resolve namespace prefixes on its own

<gpilz> http://lists.w3.org/Archives/Public/public-ws-resource-access/2010Jan/0130.html

<DaveS> HHHH /a\[namespace-uri()='http://www.example.com']/b\[namespace-uri()='http://www.example.com']

<gpilz> <wsf:Expression Language="http://www.w3.org/2009/09/ws-fra/XPath-Level-1"> /a\[namespace-uri()='http://www.example.com']/b\[namespace-uri()='http://www.example.com'] </wsf:Expression>

<Yves> //*[namespace-uri()='http://foo.example.com']

<Tom_Rutt> what about eventing filters

<DaveS> Gil is doing a proposal that address three spec. Enum, Eventing, and Frag

- DRAFT -

Web Services Resource Access Working Group Teleconference

27 Jan 2010

Attendees

Contents

8205

8290

8292

8294

8180, 8299, 8302

6435

7986

7791

8191

8196

8229

8196

Summary of Action Items

Scribe.perl diagnostic output