Requirement

Guidance

Need to cover guidance on how best to use each of the options e.g. is it out of scope or can I raise a CR to enlarge the scope. 

Use cases for "no correspondence" include support for partial mappings and where the information definition is not understood.

Semantic correspondence is about documenting the relationship between a concepts terms and definitions.

Explain the process. Put emphasis on the fact that concepts are being analysed. These are “named things”, not structural elements such as xsd:choice, xsd:sequence. There is no need to look at structural elements. No need to consider syntax.

semantic correspondence between textual definitions of the OUA and AIRM

Image Removed

The specification allows four options as visualised below. Tidy up diagram as it has six options.

Image Removed

...

The specification gives four types of semantic correspondence. The first of these is the best outcome of the semantic correspondence analysis:

• a mapping from a concept in the information definition to a concept or concepts in the AIRM.
This is the best outcome of the semantic correspondence analysis. It shows that there is an equivalent or wider/narrower relationship between the concepts.

However exceptions are available

...

. The specification also allows:

...

• a declaration that the concept in the information definition is out-of-scope of the AIRM. This gives a clear statement that the concept being mapped is not found in the AIRM and that it will not be in the AIRM.

...

• For example, the concept is:
  
  • not an ATM related piece of information

...

...

• • technology-specific or protocol-specific data

...

...

• • only a container of other ATM data

...

• • for the purpose of a specific data exchange

...

• a reference to a change request for the AIRM that intends to change the AIRM to cover the concept from the information definition. This is used when a gap has been identified in the AIRM and a request has been made to fill the gap. This obviously covers concepts that are new or have not yet been
identified
• captured but which are in the AIRM’s scope.
• a declaration that no semantic correspondence has been established for the concept.  This is used when none of the other outcomes are correct. It could, for example, be that the information definition itself is not clear enough to be able to use any of the other three options. Obviously, this should be used with care. It means that a user cannot be sure about the semantics of this un-mapped concept.

The semantic correspondence document may contain a mixture of the four types. Each of the information definition's concepts must have a semantic correspondence. For example, the figure below contains six mappings. It shows how each concept in the information definition (to the left) has a mapping.

Image Added

The specification does not set out a method for choosing the correct option. Sometimes the semantic correspondence author will need to chose the correct option with care.

...

Explain that they are a cascadeSomethings should be kept in mind when doing this:

• remember that concepts are being analysed, not structural elements such as xsd:choice or xsd:sequence. There is no need to look at structural elements.
• the textual definitions of concepts are analysed.

• there is no need to consider syntax.

  
  

Verification Support

Completeness

...

Check that the :

[  ] Each semantic correspondence

...

statements contain one of the four options

...

allowed by the specification.

Examples

This example shows how to embed a mapping into an XML schema document using the annotation element. It shows how the trace (or multiple traces) should use the AIRM's unique identifier as required by SWIM-INFO-019. Further mapping examples are available on the Understanding and recording mappings page.

<xs:annotation> 
 <xs:documentation>  
  <semanticCorrespondence>   
   <mapping>    
<trace> 
    <trace type="informationConceptTrace">-AIRM unique identifier-</trace>
    ...
    <trace type="narrowingTrace">-AIRM unique identifier-</trace>   
   </mapping>  
  </semanticCorrespondence> 
 </xs:documentation>
</xs:annotation> 

This example shows how to embed an out-of-scope declaration into an XML schema document using the annotation element. It shows how the rationale, required by SWIM-INFO-015, is added as an attribute.

<xs:annotation> 
 <xs:documentation>  
  <semanticCorrespondence>   
   <outOfScope rationale="container"/>  
  </semanticCorrespondence> 
 </xs:documentation>
</xs:annotation> 

This example shows how to embed a reference to a change request into an XML schema document using the annotation element. It shows how the change request number is captured as an attribute.

<xs:annotation> 
 <xs:documentation>  
  <semanticCorrespondence>   
   <changeRequest number="100"/>  
  </semanticCorrespondence> 
 </xs:documentation>
</xs:annotation> 

This example shows how to embed a no semantic correspondence established declaration into an XML schema document using the annotation element. It shows how the rationale, required by SWIM-INFO-015, is added as an attribute.

<xs:annotation> 
 <xs:documentation>  
  <semanticCorrespondence>  
   <noSemanticCorrespondence rationale="incomplete mapping"/>  
  </semanticCorrespondence> 
 </xs:documentation>
</xs:annotation>

This example shows how to embed a mapping into an XML schema document using the annotation element. It builds on the other examples and includes the <xs:appinfo> element that specifies information to be used by the application. This is an optional element and can be used when the intention is to use an application to further process the semantic correspondences.

<xs:annotation> 
 <xs:appinfo>Semantic Correspondence</xs:appinfo>
 <xs:documentation>  
  <semanticCorrespondence>   
   <mapping>    
    <trace type="work in progress"/> "informationConceptTrace">-AIRM unique identifier-</trace>
    ...
    <trace type="narrowingTrace">-AIRM unique identifier-</trace>   
   </mapping>  
  </semanticCorrespondence> 
 </xs:documentation>
</xs:annotation>