31 lines
1.8 KiB
Plaintext
31 lines
1.8 KiB
Plaintext
---++ Use Case: Referencing Schema Elements
|
|
|
|
----
|
|
---+++++ Description
|
|
One possible approach to developing an ontological framework for biodiversity data exchange would be to register a set of identifiers for standard data classes (Specimen, TaxonName, etc.) so that data providers can assert more clearly what types of data they wish to share. Each such class would have some kind of GUID to identify it.
|
|
|
|
In addition, similar identifiers may be assigned to each of the attributes associated with these classes (CollectionDate, PublicationYear, etc.).
|
|
|
|
If all of these identifiers are managed in an ontology independently of specific representational shemas (such as Darwin Core or ABCD), they can provide a framework onto which those schemas may be mapped. This would lead to a secure basis for interchange between different formats.
|
|
|
|
Note too that many of the attributes currently represented by schema elements in a representation such as Darwin Core could be normalised in such a model. Instead of each specimen record including attributes such as ContinentOrOcean, Country, StateOrProvince, County, Locality, Latitude and Longitude, the Specimen object might have a single attribute LocalityId that held a GUID identifying a different object of the class Locality. (Note that the Locality class would then have a GUID identifying the class, and that each Locality object would have its own GUID).
|
|
|
|
----
|
|
|
|
---+++++ Categories
|
|
CategoryUseCases
|
|
|
|
|
|
---++ Comments
|
|
Use the space below to make comments about this page.
|
|
|
|
------
|
|
|
|
%ICON{bubble}% Posted by Main.RobertHuber - 2006-01-18 11:07:34
|
|
|
|
Interesting approach, however this might lead to a GUID overkill?<br />
|
|
GUIDs for 'top' classes (Specimen, TaxonName, Locality, Sample...) yes, but assigning them for simple attributes probably goes too far?
|
|
|
|
------
|
|
|
|
%ICON{bubble}% |