[ASN.1] Inheritance of the extension marker - H.450.1 Question

[frank . derks]

Over the last couple of weeks there have been discussions in several mailinglists 
on the interpretation of some of the ASN.1 in H.450.1. One of the discussions is 
about the "inheritance of extension markers".

The following definitions are from H.450.1:

InvokeIdSet INTEGER ::= {InvokeIDs,...}
InvokeIDs ::=   INTEGER (0..65535)
InvokeId ::= INTEGER

Invoke {InvokeId:InvokeIdSet, OPERATION:Operations} ::= SEQUENCE
{
  invokeId  InvokeId (InvokeIdSet)
            (CONSTRAINED BY {-- must be unambiguous --}
            ! RejectProblem : invoke-duplicateInvocation),
  linkedId  InvokeId
            (CONSTRAINED BY {-- must identify an outstanding operation --}
            ! RejectProblem : invoke-unrecognizedLinkedId)
            (CONSTRAINED BY {-- which has one or more linked operations--}
            ! RejectProblem : invoke-linkedResponseUnexpected) 
            OPTIONAL,
  opcode    OPERATION.&operationCode
            ({Operations}
            ! RejectProblem : invoke-unrecognizedOperation), 
  argument  OPERATION.&ArgumentType
            ({Operations} {@opcode}
            ! RejectProblem : invoke-mistypedArgument) 
            OPTIONAL
}

In the above, invokeId is defined as being of the type InvokeId, but constrained
by InvokeIdSet. InvokeIdSet, however has an extension marker in its definition
and it is not clear whether this extension marker is inherited by the definition
of invokeId. I.e. could the definition of invokedId be replaced by

  a) "invokeId   INTEGER (0..65535)"

or by

  b) "invokeId  INTEGER (0..65535, ...)"


47.4/X.680 (12/97 version) states:

"If a type defined with an extensible constraint is referenced in a 
"ContainedSubtype", the newly defined type does not inherit the extension 
marker or any of its extension additions.  If the newly defined type is meant 
to be extensible, then an extension marker shall be explicitly added to its 
"ElementSetSpecs"." 

Applying this rule to the definition of invokeId would now first seem to 
raise the question as to whether this definition should be treated as 
a type definition or not. If so, the extension marker is not inherited. 
Otherwise, the situation is still unclear.

Regards,

Frank Derks