relationships

[dustymc]

From #8384 - do we need some ability to do something that we can't do now?

Suggest starting with functional needs, not models or possible implementations.

Current CT: https://arctos.database.museum/info/ctDocumentation.cfm?table=ctid_references

[Jegelewicz]

GloBI is a public repository that aggregates biotic relationships. GloBI relies on the OBO Relation Ontology. Why not follow what GloBI does for biological interactions?

There are plenty of relationships where the related object/specimen is not cataloged, so we have also recorded those in a multitude of attributes with no taxonomic control. I think that should all be considered as part of this as well, because finding "relationships" in Arctos requires searching in many places (identifications, attributes, identifiers).

[dustymc]

Why not follow what GloBI does

It's a very different model - which doesn't mean we can't do nearly anything, as long as we keep our reality in sight.

OBO Relation Ontology.

There's one obstacle: Adding some meteorite-or-whatever relationship would, under a 'just use the ontology' model, first require adding that to some ontology and/or adding that ontology to the mashup. Arctos scales broader than almost anything else. (This still absolutely doesn't mean we can't do something awesome, just that being awesome-r would be a bit more work at a bit different scale than we're used to.)

related object/specimen is not cataloged

That's the core difference, Arctos is not so simple as the semantic triple that everyone in ontology-land expects, and that may or may not be an obstacle, and it may or may not be surmountable if it is.

https://github.com/ArctosDB/BackBurner/issues/25 - there's clearly an ontology community, there's clearly interest in joining it, someone write the proposal if that's where The Community wants to go!