SAWSDL WG teleconference
18 Jul 2006

Approval of last week's minutes

-> http://www.w3.org/2002/ws/sawsdl/minutes/20060711

Carine: Approval of minutes, any comments? no comments, minutes are approved

Action Items review

<scribe> ACTION: Terry to review last call of WSDL RDF mapping [DROPPED] [recorded in http://www.w3.org/2006/07/18-semann-minutes.html#action01]

<scribe> ACTION: EricP to submit our review to WSDL RDF mapping to the WSDL WG [DONE [recorded in http://www.w3.org/2006/07/18-semann-minutes.html#action02]

-> http://www.w3.org/2002/09/wbs/39001/200608torino/

<scribe> ACTION: Rama to ping Laura about sending the use case to the Group [recorded in http://www.w3.org/2006/07/18-semann-minutes.htm#action03]

Administrativia

Carine: registration for the next F2F meeting is open, only 2 people are registered, please register!

Documents status

Carine: Example Document Status

Rama: Examples need more work, also section for best practicies
Rama: I will continue to work on it and the improved version should be ready for the review before next F2F meeting

Carine: Do you want more input?

Rama: Telecom Italia sent input - I have to read it and see how it fits
Rama: Document will be ready for the review after changes will be incorporated by end of July

Action: Rama to ask Laura to send the Telecom Italia UC document to the group

Rama: Ok

Tomas: Should this UC from Telecom Italia be for more elaborated use cases from particular domain?
Tomas: In the existing example document, there are only "simple" use cases not related to the domain

Rama: Document is big, could serve as a kind of example in appendix
Rama: I need to read it and will see what makes sense to use - which part to put to the main doc and which to appendix

Definition of semantics

(from Rama)

Rama: The idea was to deinfe semantics as meaning of objects which could be controversial
Rama: Better to define it as meaning of concepts in a domain model and relations between concepts
Rama: Carlos was pointing out that we should reorganize the definitions and not repeat them
Rama: Ok to define terms as long as they are defined properly

Joel: Carlos suggested to use semantic model instead of domain model
Joel: Domain model is used in a larger context (e.g. telecom, etc.)
Joel: We can use domain model as a kind of semantic model

Rama: ok

Laurent: It is overspecified, there is no mention of context in spec, it is beyond the scope

Rama: Context is defined as relation between concepts

Joel: Are we changing order of terms or not?

John: First define domain model and then semantics

Rama: Yes
Rama: Instead of "domain model" we say "domain"

All: agree

Carine: There is no place where domain model is used somewhere else?

Rama: When it is used for the first time it should be defined

Laurent: Do we need a definition of a domain?

John: This could be never ending

Rama: Joel write it up and we get back to it

Action: Joel to update definitions in the document

Issue no 22: revisit section on WSDL 1.1 mapping

John: switch from WSDL 1.1 to 1.2 as WSDL 1.1 does not allow extensiblity of operation

Carine: 1.1 is there to show how we deal with legacy services. 1.2 does not exist (it became 2.0)

Rama: In 1.1 operation is not extensible, doc annotation tag used to extend operation

Joel: We put at least a fragment example - should WSDL1.1 example be in the main doc or in the appendix?
Joel: will be less then a page

Carine: This issue will become editorial - keep it open and we get back to it?

Joel: Leave it open until the next call

Issue no 8: categorization

Rama: interface can have modelReference, they can be used to specify categorization
Rama: interface can have multiple operations
Rama: why operation can be classified, why to classify interface if you have operation
Rama: we don't discuss how operations could be categorized
Rama: will category for operation override category of interface?

John: will there be example to clarify this? Could be informative

Laurent: There could be also description of choreography, non-functional properties,
Laurent: preconditions, postconditions attached to the operation with dependencies
Laurent: There is no need for preference for modelRoferences

Carine: We should not go too deeply to define use of modelReferences. We don't need to clarify whatever modelReference can contain. We already resolved that all modelReferences apply.

Carine: Should the categorization example clarify how do we deal with multiple modelReferences for categorization on interface and operations?

Rama: If we talk about categories, it must be clarified how to categorize
Rama: we need to give advice/suggestion how to use modelReferences on interface and operation

John: Should not be this clarified in best practices?

Rama: Yes. We can refer to the use case

Action: Rama to come up with a text explaining "best practice" for the categorization issue