issue 88 - notes from discussion on Aug 10 meeting

[Diane Jordan <drj@us.ibm.com>]

This is to document the discussion we held last week
and this week on Paco's latest proposal for 88 so we can proceed when he
returns:

• Three points on transitivity:
  • Verify Paco agrees with the change proposed by Yaron:
  Change:
  
  However, documents (or namespaces) imported by an imported document (or
  namespace) are no considered to be transitively imported by the process.
  
  To:
  
  However, documents (or namespaces) imported by an imported document (or
  namespace) are not considered to be transitively imported by the process.
  
  The lack of transitivity can cause some interesting issues. For example,
  a document D1 could define a type called a:Type. But a:Type's definition
  could depend on another type called another:Type which is defined in
  document D2. D1 could include an import for D2 thus making
  another:Type's definition available for use with a:Type.
  
  Now imagine a BPEL process refers to a:Type and imports one and only one
  document, D1. By importing D1 the BPEL process can legally refer to
  a:Type. But, the BPEL process could not refer to another:Type even
  though D1 imports D2. This is because transitivity of import is not
  supported by BPEL. Note, however, that D1 can still import D2 and a:Type
  can still use another:Type in its definition but the BPEL process cannot
  directly refer to another:Type. In order to allow the BPEL process to
  refer to another:Type it would be necessary for the BPEL process to
  directly import document D2.
  
  • Suggestion from Monica to add normitive wording on how transivitity should work:  
  Schema definitions defined in the types section of a WSDL document
  >> which is
  >> imported by a BPEL process definition are considered to be effectively
  >> imported themselves and are available to the process for the purpose of
  >> defining XML Schema variables. However, documents (or namespaces)
  >> imported
  >> by an imported document (or namespace) are no considered to be
  >> transitively
  >> imported by the process.
  >
  > mm1: Paco, shouldn't we add some 'teeth' to the last statement, such as:
  >
  >    "Documents (or namespaces) imported by an imported document (or
  >    namespace) MUST NOT be transitively imported into the process. "
  
  • Address question from Alex on relationship between transitive and non-transitive closure and potential conflicts
  • No Namespace defined:  Prasad still doesn't understand but is ok with proceeding.  This raised the question of whether there should be more explicit clarification on what the engine should do.  
  • Chris made comment:  We may need to change the property alias. Property alias is not directly referenced by BPEL process. Property alias is not referred to or not really referenced. Should we require including property alias? Property alias is not a name.
  Update at Aug 17 meeting from Chris - this is mostly a wsdl question and he is op with the property alias reference in the proposal.  
  
  
  Regards, Diane
  IBM  Emerging Internet Software Standards
  drj@us.ibm.com
  (919)254-7221 or 8-444-7221, Mobile: 919-624-5123, Fax 845-491-5709