Requirement
Title | Mapping of data concepts to the matching AIRM concepts |
Identifier | SWIM-INFO-017 |
Requirement | The mapping of a data concept shall contain a trace from the data concept in the information definition to the AIRM concept that has an equivalent or wider meaning and a trace to the data type in the AIRM that has an equivalent or wider meaning. |
Rationale | This includes a basic trace to establish the semantic correspondence between concepts. Without such a trace the mapping is pointless. The requirement adds a second trace for data concepts, as the constraint on value space expressed by the data type is also important to the semantics. |
Verification | Correctness |
Examples/Notes | Example: An example is an information definition contains a data concept called “Target Startup Approval Time” that is a time. Two traces are needed for this mapping:
Note: In cases where the trace is to an AIRM concept that has a wider meaning, SWIM-INFO-018 also applies. An example of a trace to an AIRM concept with a wider meaning can be found in SWIM-INFO-018. Note: It may be possible to combine these traces into one single statement depending on the tracing techniques adopted. Note: This requirement implies that implementations using technology that provides no typing (for example, JSON) will need to specify the applicable value range constraints at design time relative to the AIRM data type to demonstrate semantic correspondence. Note: SWIM-INFO-019 requires the use of the AIRM’s unique identifiers in traces. |
Level of Implementation | Mandatory |
Guidance
This requirement covers the mapping of data concepts. This requires two traces.
The first trace is similar to that explained in SWIM-INFO-016 Mapping of information concepts.
The second trace is particular to data concepts. This trace still concerns the semantics of the data types being mapped to each other. However, it is important to bear in mind the requirement on the compatibility of data types - that requirement will make data concept traces much easier.
It is usually easy to spot exact matches when tracing data types. The skill is when they do not match exactly. Some things then need to be considered.
- The specification talks about an "equivalent" meaning. This allows some differences in the data types.
- The requirement on the compatibility of data types gives good examples of there the data types can be considered as equivalent.
- The specification talks about a "wider" meaning.
- The requirement on the compatibility of data types explains that the compatibility of data types can be achieved in several ways
- by "casting" of data types. This, for example, means that a numeric data type can always be regarded as a string. This means a trace can be established when the information definition's data type can be cast to the AIRM data type.
- likewise a string of length 4 can always "fit" a string of length 32.
- by "casting" of data types. This, for example, means that a numeric data type can always be regarded as a string. This means a trace can be established when the information definition's data type can be cast to the AIRM data type.
- The requirement on the compatibility of data types explains that the compatibility of data types can be achieved in several ways
It is also possible to rule out some traces:
- When the information definition's data type has a wider, more general, meaning no trace can be made. This may mean:
- The data concept mapping is wrong. In effect, the data concept is not the same as the AIRM concept.
- A change request may be considered if the AIRM data type is wrong.
- The data concept mapping is wrong. In effect, the data concept is not the same as the AIRM concept.
Examples
- simple mapping e.g. characterString and numeric
- complex mapping such as estimated take-off time
- geometries as a way to better explore compatibility.
How to read the AIRM to find the data types etc. Maybe an annex to the supporting material. Where should I start e,g, LM or IM.
- Information concepts are in the IM.
- Data concepts and their Datatypes are in the LM
- Basic type semantics are in the Foundation.
Two use cases - place at the root of this semantic correspondence section:
- technical - near service developers - use LM
- operational - e.g. for information exchange requirements - use IM.
Other notes
- if you have a data type but the AIRM does not you are probably in the wrong AIRM part. Consider looking at the LM.
- add more on the compatibility story.
SEE COMPLIANCE HANDBOOK FOR IDEAS
Verification Support
Correctness
Check that:
- Each trace from the data concept in the information definition to the AIRM concept that has an equivalent or wider meaning is correct.
- Each trace to the data type in the AIRM that has an equivalent or wider meaning is correct.
Examples/Best Practice
<gmd:title> guess what </gmd:title>