Closed zhengj2007 closed 5 years ago
You mean rdfs:isDefinedBy https://www.w3.org/TR/rdf-schema/#ch_isdefinedby
Yes. It is rdfs:isDefinedBy
So this could be used to hand a term's curation from one ontology to another, while maintaining its purl/ID? A way of tracing its provenance? Then services like OntoBee could use rdfs:isDefinedBy to enable the term to be displayed in the context of the adopting ontology?
This is possible, but it complicates a lot of infrastructure. Would this be implemented in just ontobee - what about OLS? What is the mechanism to avoid me injecting an exploit to steal terms? E.g. I could cause you problems by accidentally or maliciously injecting FOODON_nnnn isDefinedBy UBERON in UBERON.
There is a trade-off here. In some cases the minimal churn caused by minting a new ID is preferable.
I see your point. Yes it would have to be a standard that ontology consumers buy into, operating alongside the term deprecated / replaced by mechanism.
(I know its too late but the alternate universe would have a simple TERM_XYZ factory that all ontologies would subscribe to, and rdfs:isDefinedBy would point to the curating ontology of each term).
We currently have a small number of overrides defined in the PURL system: https://github.com/OBOFoundry/purl.obolibrary.org/blob/master/config/obo.yml#L22 Building that into our central authority addresses some of the points that Chris raised. The implementation could be more elegant, and we could make it more useful for the portals.
Discussed on 2018-12-4 OFOC call.
The annotation property, rdfs:isDefinedBy, is used to indicate a term with external ontology prefix is maintained by current ontology, not an imported term.
Some BFO prefix relations are in RO. Since it with BFO prefix, it caused problem when software used the term prefix to look up the terms. see issue: https://github.com/BFO-ontology/BFO/issues/218
Besides, OBI defined some high level terms that may belong to other ontologies, such as OGMS, IDO. If we can specify the owner rather than prefix. It allows suitable ontology to take over the terms without deprecation and assigning new ID.
I remembered IAO:is defined by has been proposed to use for this purpose. Any decision on it or other suggestion?