Re: Deriving various simpleContent and complexContent types from the same base?

"Henry S. Thompson" wrote:
> 
> Simon Cox <Simon.Cox@dem.csiro.au> writes:
> 
> > 1.  I want to define an abstract element <_Observation>,
> > whose substitutionGroup includes elements such as
> >   <Quantity> which has simple content of type xsd:double
> >   <Value> with simple content of type xsd:string
> >   <Count> with simple content of type xsd:nonNegativeInteger
> >
> > e.g. Can I define an abstract base type with no content
> > (complexType?), or of "anyType" (can this be a simpleType?),
> > and then "restrict" the base to the various simpleTypes?
> 
> Yes.

How?


Here's my current workarounds, with comments.  
It leads to me having to use <xsd:group ref="qf:Observation"> in content models, 
where I prefer to see <xsd:element ref="qf:_Observation">.


  <!-- Generic (abstract) Complex Observation -->
  <xsd:element name="_ComplexObservation" type="qf:AbstractComplexObservationType"
abstract="true"/>
  <!-- AbstractComplexObservationType is an empty complex type 
    -- to which specific content models can be inserted by extension -->

  <!-- Generic Scalar Observations -->
  <xsd:element name="Count" type="xsd:nonNegativeInteger"/>
  <xsd:element name="Quantity" type="qf:quant"/>
  <xsd:element name="Value" type="xsd:string"/>
  <xsd:group name="ScalarObservation">
	<xsd:choice>
		<xsd:element ref="qf:Value"/>
		<xsd:element ref="qf:Count"/>
		<xsd:element ref="qf:Quantity"/>
	</xsd:choice>
  </xsd:group>
  <!-- not really happy about using a choice-group here, 
    -- would prefer an abstract element and substitution group, 
    -- but simpleType elements cannot be declared "abstract" 
    -- and also I can't figure out how to derive specific 
    -- simple types by restriction from anyType -->

  <!-- The general Observation group -->
  <xsd:group name="Observation">
	<xsd:choice>
		<xsd:element ref="qf:_ComplexObservation"/>
		<xsd:group ref="qf:ScalarObservation"/>
	</xsd:choice>
  </xsd:group>
  <!-- not really happy about using a choice-group here, 
    -- would prefer an abstract element and substitution group, 
    -- but can't figure out how to derive both simple and complex types from anyType -->



[Is this the right list for this kind of inquiry?]
-- 
Simon.Cox@dem.csiro.au    CSIRO Exploration & Mining
T:+61(8) 9284 8443 F:+61(8) 9389 1906 M:0403 302 672
http://www.ned.dem.csiro.au/research/visualisation/

Received on Wednesday, 21 February 2001 21:46:47 UTC