See also: IRC log
<scribe> scribe: Carine
<scribe> scribeNick: carib-off
<JacekK> http://www.w3.org/2002/ws/sawsdl/minutes/20060606
RESOLUTION: Minutes are approved
<scribe> ACTION: Terry to review last call of WSDL RDF mapping http://www.w3.org/TR/wsdl20-rdf/ by beginning of July [PENDING] [recorded in http://www.w3.org/2002/ws/sawsdl/minutes/20060613#action01]
<scribe> ACTION: EricP to review last call of WSDL RDF mapping http://www.w3.org/TR/wsdl20-rdf/ by beginning of July [PENDING] [recorded in http://www.w3.org/2002/ws/sawsdl/minutes/20060613#action02]
<scribe> ACTION: JacekK to open 2 issues: relation of multiple model references (1) and annotationContext (2) [DONE] [recorded in http://www.w3.org/2002/ws/sawsdl/minutes/20060613#action03]
<scribe> ACTION: LaurentH to specify use cases for separating different annotations (for example optimizations) into different contexts/perspectives [DROPPED] [recorded in http://www.w3.org/2002/ws/sawsdl/minutes/20060613#action04]
LaurentH: I accept that we loose the info because it can be found in references
... we don't need it inside the WSDL itself
<scribe> ACTION: JohnMiller to propose a specific RDFS ontology for interface categories [DONE] [recorded in http://www.w3.org/2002/ws/sawsdl/minutes/20060613#action05]
JK: logistics page http://www.w3.org/2002/ws/sawsdl/admin/meetings/200606Galway
... it means no telcon next week
... we need to freeze the documents ASAP so that we can all read the same version before the F2F
JF: I think I can get the latest changes soon
... friday
JK: let's add any changes we can before tomorrow COB, then freeze
... After freezing the document, please send an email to the WG
<JacekK> fixed email URI http://lists.w3.org/Archives/Public/public-ws-semann/2006Jun/0044
JK: proposal is to close issue 18 with implicit conjunction and if conflict, document is invalid
John: Union or Intersection?
JK: it would not be used for subsumption but inconsistency checking
LH: You said modelReference was very vague. I think a tool could decide to drop a modelReference, it's not a problem of consistency
... we just put things together. Conjunction is too close to ontologies languages
RA: implicit conjunction is dangerous, we may open problems we can't foresee
... we'd better not say anything
JK: I think I agree with Rama that saying we don't specify relationship between modelRefs is an option
John: the best solution would be probably to have only one modelRef but it would be too draconian
Amit: we should encourage people to annotate at the higher level (i.e. not annotate Full name with firstname/lastname)
RA: we may need Best Practices guidance document
EricP: if several authors add modelRefs in the SAWSDL (e.g. for different logics), I think we're stuck saying we don't guarantee there's no connection
LH: if you can refer to an aggregation of ontologies, it would be a better solution
<laurenth> using an aggregation should not imply to define the proper ontology, although it can be in the best practices
RESOLUTION: Examples document will contain Best Practices documenting proper uses
<Amit-John> Proposed rolution: single modelRefs encouraged, ensuring ontology model what is needed for this is a better way then addressing the deficiency through multiple modelRefs
JK: I propose to close issue 18 with no action
RA: if we can't fix the ontologies, it may not such a bad thing to use (first name, last name) in the annotation
... I agree that fixing the ontology is the right thing to do, but in real world it's not uncommon to have multiple levels of description
JK: it looks like it's material for the BP
JF: all modelRefs *apply* but not specify any relationship between them
LH: yes, say nothing about inconsistencies (no logical relationship)
RESOLUTION: issue 18 closed, we will not specify any relationship between the modelReferences
<JacekK> ACTION: Examples document editors to add "best practices" section, and to add the firstName/lastName scenario description [recorded in http://www.w3.org/2002/ws/sawsdl/minutes/20060613#action06]
JK: Laurent proposed to drop it
LH: I suggested in an email to move modelReference into an element, which could prepare for future evolution (more info in modelRef)
... including annotation context
... why is it so difficult to change from attribute to element?
JK: I was planning to open another issue for this
... it will be discussed at the F2F
... on issue 17 itself, any ojection to close it with no action?
RESOLUTION: issue 17 closed with No Action
JK: do we need to transform in both directions?
AS: bidirectional mapping is too complex
RA: if we allow multiple schemaMapping, again we don't have to specify how to use them
... we could have BP on this
JK: a specific language to map XML to ontologies could be bidirectional
... XSLT really goes in 1 direction, we could have 2 xslts
EricP: I think we really want to label in which direction we want to go
... to know how it works
JK: in the case of schemaMappings, we know that there are exactly 2 directions
... unlike the modelRef problem
... we could split schemaMapping in 2 attributes
... for the 2 directions
RA: all the directions, target, language... information could be in the external file
EricP: if it was more complex than just up and down, I'd support external info
JK: upcast/downcasting vs. lowering/lifting language
<Amit-John> upcasting...
<Amit-John> since it is programming language related
<Amit-John> ok weith Laurent's proposal
JK: objections to adopting lowering/lifting names?
RESOLUTION: SchemaMapping is split into 2 attributes, loweringSchemaMapping and liftingSchemaMapping
JK: the rest of the issue (#6) will be discussed later.
... Adjourned.