investigate variables as IRIs about owlapi HOT 4 OPEN

SWRLRule in Manchester Syntax use ?x for variables now

from owlapi.

I think they are IRIs in the original SWRL "spec".

from owlapi.

I've added the workaround that Protege uses in the parser so that ?x is equivalent to ?urn:swrl#x.
For Protege, as far as I understand it, a rule in Manchester syntax should always have variables of the form ?foo, and in the object model they always look like urn:swrl#foo.
I've changed the manchester syntax renderer to output ?foo rather than ?<whatevernamespace#foo>, but I'm now wondering if this should always be the case or if there should be a parameter, e.g., in the format, to control this behaviour.

from owlapi.

Changed this to a more conservative approach. If there is a namespace different from urn:swrl#, use the old full IRI rendering, otherwise use the simplified form. The reverse for parsing.
This allows for simple syntax for those rules where the namespace is irrelevant, and keeps the namespace in other cases (if there is any other case, or not). Ontologies with rules roundtrip correctly with either format.

It's a pity that the SWRLVariable name implies SWRL is involved, since the interface does not have anything tying it specifically to SWRL. Might introduce a Variable interface in version 4

from owlapi.