Data mediators are similar to topic maps. They take heterogeneous data and present a common interface to it.
Does that make a data mediator mapping a topic map?
No!
Here is a test to see if a data mediator mapping qualifies as a topic map:
- Create two mappings of the same data using mediators that use different vocabularies.
- State what basis you would combine components from the two mappings. (Or even the basis for mapping from the data source to the target, but I digress.)
What a data mediator creates is a useful but blind mapping. That is the reason for that mapping is not apparent from the map.
That prevents re-use of the mapping by others. Such as combining it with other mappings.
The reason for the mapping, subject identification in topic map terms, remains in the head of the person who created the mapping.
What happens when they move up, on or simply retire?
The business case question is:
How many times have you paid to have subjects in your information system identified?
Or even better:
How many more times are you going to pay to have subjects in your information system identified?
Is this time going to be the last time? Could be if you were using topic maps.