As I mentioned in Maps and Territories, relational database theory excludes SQL schemas from the items that can be modeled/mapped by a relational database.
All maps are lossy, but I think we can distinguish between types of loss.
Some losses are voluntary, in the sense that we choose, due to lack of interest, funding, fitness for use, or other reason to exclude some things from a map.
We could in a library catalog, which is a map of the library’s holdings, add the number of words on each page of each item to that map. Or not. But that would be a voluntary choice on our part.
The exclusion of SQL schemas from the mappings possible within the relational database paradigm, strikes me as a different type of loss. That is an involuntary loss that is mandated by the paradigm.
It simply isn’t possible to model an SQL schema in the relational paradigm. Those subjects, the subjects of the schema, are simply off limits to everyone writing an SQL schema.
I mention that because with topic maps, all the losses are voluntary. At least in the sense that the paradigm does not mandate the exclusion of any subjects, although particular legends may.
I think it would be helpful to have a table listing model/mapping systems and what, if anything, they exclude from modeling/mapping.
Suggestions?