Minutes for twenty-sixth OGSI-WG Teleconference
February 17, 2003 @ 17:30 - 19:30 GMT

Nick Butler - IBM
Dave Snelling - Fujitsu
Pete Vanderbilt - NASA
Steve Graham - IBM
Steve Tuecke - ANL
Karl Czajkowski - ISI 

1) Approve minutes of February 15-16th Teleconference - OK

2) Matters Arising:

2a) Resolve how to express extensibility parameters

   There was a request for a qname to the operation. It appears that this is
   currently not allowed in WSDL. Also an output tied to the input, but this 
   is the start of the slipper slope to reinventing WSDL.

      Resolved: Leave it as it is, i.e. Steve T's proposal of Feb 16.

   What about a no argument?

      Rejected: Declare the argument type as ogsi:nil.
      Resolved: Make the inputElement attribute optional and interpret this as 
                allowing no argument.

2b) Resolve whether to use fault message part type= or element=

   Proposal: Go with element= based on the WS Interop recommendation.

2c) The name of the draft

   Resolved: Remove.

2d) Fault codes

   Resolved: Leave the fault. Add to the fault code an attribute 'faultscheme' indicating
             the namespace for the interpretation of the fault code scheme.

2e) In Resolver exclusion set, what does this mean?

   The GSRExclusionSet MAY contain any number of GSHs, but they MUST be ignored
   by the resolver.

   Resolved: No real problem here, but the words could be clearer. Also remove
             'hint' and change 'should' to 'SHOULD'.

2f) Fault names

   How do we get the name of a fault from the faults contained in the cause.

   Action: Note in the text that xsiType should be employed in the cause section of a Fault.

2g) Member interface argument in Service Group

   Should we include the memberInterfaces as an argument to the add operation.

   Resloved: Leave it as it is.

   Resolved: In 13.2.1 model language from factoryLocator in GridService 
             for memberserviceLocator.

2h) Comments on ServiceGroup

   Resolved: In 13.1:

     Kill: There SHOULD be only one Entry per member Grid service instance.

     Extend: Grid service instance MAY be included in a ServiceGroup multiple times.
             With words 1) It is a bag. 2) Designers MAY subclass to set semantics.


     Resolved: Upper case membershipContentRuleType.

     Resolved: Explain membershipContentRuleType a bit more.

     Resolved: Nuke memberInterface in entry, serviceGroupEnrty, and
               ActualContentArgument e.g. Use EntryContentType.

     Resolved: Change to MUST in: It also contains the set of element types
               (content elements) that SHOULD be included in the content for
               the given portType. Plus it MAY include further content.

     Resolved: Add words to the effect that if more that one rule applies, all
               MUST be satisfied. Also, aliasing of instances is not an issue
               in the content rule, but all portTypes are checked against the rules.

     Action: Dave S 22 -> 23

3) Action Review:

Action:  86: Feb 12: Steve T: Check that our service data elements that need coherency semantics, e.g. non-static/non-constant, have them specified. - Done (2/16)

Action: Execute Steve T's proposal as of email:  Feb 16, 2003: Extensible operations proposal - Done (2/16)

Resolved: Use converge language (see serviceGroupEntry).

Action: Steve T: Convert Service/portType to interface in the spec, other than ServiceGroup. - Done (2/16)

Action:  97: --- --: Pete V: To find cases where an alternative to emptyExtensibility matters. - Done (2/16)

4) WSDL Schema Issue:

   It has been suggested that we put the WSDL schema under CVS or the like.
      - Who gets write permission?
      - Is there a verification process possible?
      - Should it include a Notification service?
      - How long do we allow changes?

   Resolved: List the Schema at the end, call the bits in the text non-normative.

   Resolved: Put a reference into the final chapter, to be inserted later.

   Resolved: Nuke the change logs.

9) Next Meeting: 

As required (ideally GGF).