ELK 0.4.3 claims inconsistency for ontologies without individuals.

[psiotwo]

With input file input.ttl:

@prefix : <http://test.org/> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .

owl:Thing rdfs:subClassOf owl:Nothing .

and the command:

robot reason -i input.ttl --reasoner ELK -o output.ttl

I am getting "ERROR The ontology is inconsistent. TIP: use a tool like Protege to find explanations".

This particular ontology is an edge case, but the same problem can easily manifest (for example) when this the unsatisfiability of owl:Thing is achieved through an incorrect interplay of domains/ranges and disjointWith axioms (which is actually where I started tracking this issue ... ). So heavily in favour of merging the ELK 0.5.0 support PR.

Checked with standalone ELK 0.4.3 and the current 0.5.0-SNAPSHOT @ Github and it seems to be resolved in the 0.5.0.

[matentzn]

@psiotwo Not sure I understand - this ontology is by definition inconsistent. If Elk 0.5 does not detect this case, then ELk 0.5 is incomplete!

[psiotwo]

Hm, sure you are right - I forgot the int. domain must be non-empty by def and then was confused by less verbose logging of ELK 0.5.0 that didn't report inconsistency in the default log level ... shame on me!