Open dosumis opened 7 years ago
Yes, OPAs to edges is the most obvious, and is the undocumented behavior now
Not sure I follow the reification
CAs - own axiom type or an edge? I am tentatively opting for the latter
@kltm this differs slightly from ABoxes in MinervaJSON
Not sure I follow the reification
Seems clear enough to me:
ClassA has_exemplar value IndividualB <-> "subj" : "Class A", "pred" : "has_exemplar", "obj" : "Individual B"
I've occasionally found this pattern useful for VFB. I think it may become more common in future.
Can we try to keep to a single vocab for this?
OPA = FACT CA (class assertion) = type axiom ?
Would be fantastically useful for VFB if generation could be extended to cover at least the I:C and I:I mappings above.
Type statement, simple anonymous class (R some B):
"subj" : "Individual A",
"pred" : "part_of",
"obj" : "Class B"
This is a nice convenience pattern. However, the reverse mapping becomes ambiguous if we have punning. But this may not be a problem if we say punning out of scope: #11
hasValue pattern understood (it was the reificiation terminology that threw me). Same comments as above.
How should we handle negative property assertion axioms?
Is it legal in obographs to include individuals in the 'existential graph'? Could just have edge translation to OWL depend on owl entity type of object and subject:
C:C SubClassOf to name class "subj" : "Class A", "pred" : "Is_a", "obj" : "Class B"
SubClassOf to simple anonymous class (A R some B) "subj" : "Class A", "pred" : "Is_a", "obj" : "Class B"
I:C Type statement to named class: "subj" : "Individual A", "pred" : "InstanceOf", "obj" : "Class B"
Type statement, simple anonymous class (R some B): "subj" : "Individual A", "pred" : "part_of", "obj" : "Class B"
I:I FACT: "subj" : "Individual A", "pred" : "part_of", "obj" : "Individual B"
C:I Simple reification pattern (A R Value B) "subj" : "Class A", "pred" : "has_exemplar", "obj" : "Individual B"
(Edited)