I am responsible for this choice, but am happy with whatever the community decides. (if we chose a new AP we would likely make a new robot option to preserve backwards compatibility)
We can do this at different granularity:
boolean yes/no indicator
name of reasoner e.g. Elk
other metadata e.g. date at which reasoning was performed, hashes of input ontologies
We want to be able to record provenance on axioms in general. A specific case of this is recording that an axiom was generated by a reasoning process:
robot will optionally tag inferred axioms with oio:is_inferred="true":
http://robot.obolibrary.org/reason
https://github.com/ontodev/robot/blob/9cc988a56a56f23deb8df604403cfaa3484b0f53/robot-core/src/main/java/org/obolibrary/robot/ReasonOperation.java#L448-L453
I am responsible for this choice, but am happy with whatever the community decides. (if we chose a new AP we would likely make a new robot option to preserve backwards compatibility)
We can do this at different granularity:
Related: