Use existing taxonomy or thesaurus as schema

I was wondering why existing taxonomies or thesauri like the CABT or Agrovoc are not being used as schema.

This would save a lot of work, a lot of cross-referencing and xml or json expressions, while building on existing authority and distribution.


Michiel wrote:

I am still hoping for an answer on our earlier discussion on using taxonomies or thesauri as a starter for developing a practical meta data schema for farm management information.

Absence of a practical schema for farmers seems to be reflected as well  in the December 2018 report of the RDA Agrisemantics WG as distributed by the latest AIMS newsletter.

Michiel wrote:

Thank you Valeria,

I agree that my question is hypothetical with thesauri having another goal.

The origin of my question is that I am working with (smallholder) farmer groups and maintaining databases  for such groups. Taking into account the need for interoperability of datasets I was looking for a schema to use. I cannot find a proper schema available that have terms aboard like acre, hectare, pruning, milking cow etcetera.

For some other projects I maintain repositories, using the Agrovoc and the CABT as indexation service. 
As these thesauri are well established, well controlled, well authorized, I wondered why these coudn't double as metadata schema.

Maybe I should ask a different question: would it be possible to develop schema based on existing thesauri, rather than developing separate metadata schema?



valeria.pesce wrote:

CABT and AGROVOC were created as thesauri, as a set of terms to be used to index resources. Now they've evolved into concept schemes, defining a set of concepts (with basic thesaurus relationships). They don't define classes and properties. Schemas on the contrary define classes and properties to describe entities.

I am not sure if I fully understand how you would use CABT or AGROVOC as a schema. Would you use the concepts as classes? And what would you use for properties?

CABT and AGROVOC are both concept schemes with thousands of concepts. Even if you define these concepts also as classes in an RDF schema, what would you gain in comparison to having them as concepts? Would you then define properties for each of these classes?

In terms of schema functionalities, AGROVOC actually uses an additional schema (more precisely, an ontology) that defines a list of properties that can be used in addition to SKOS properties to describe relationships between the concepts (like "affects", "develops into"...). But I don't think this is what ou're looking for.

Could you give an example of a use case for an AGROVC schema?