| [higgins-dev] & does kill! |
Well, ok ... at least it seems to be a huge pain in the ass for the OWL editors I've tried. As it turns out, like Protégé, SemanticWorks also chokes on the nested entity references and will call most of the entities referenced "Unknown Resources" when doing an OWL-DL check because it parses them wrong. Once the "&" are replaced with a simple "&", all those OWL-DL errors go away and the graphical representation comes out correctly as well. So, after spending most of today trying (unsuccessfully) to figure out how to get DOM4j not to encode entity references, I'm temporarily punting with via a java.lang.String::replaceAll call. Attached is the resulting generated output where all but 2 small OWL-DL errors remain. 1. The authorityRevocationList issue caused by a bad schema on the source LDAP server and so, I'm not too concerned right now. 2. A supposed "Unknown Resource" error on my owl import statement. It's weird because it DOES find and display all the Higgins ontology and (after "fixing" the & problem) the relationships between the generated ontology and the Higgins ontology. Here's what SemanticWorks says isn't OWL-DL: Invalid OWL DL syntax in owl:imports-statement: <http://www.bandit-project.org/higgins/ontologies/2006/ldap> owl:imports <http://www.eclipse.org/higgins/ontologies/2006/higgins.owl> Unknown Resource <http://www.eclipse.org/higgins/ontologies/2006/higgins.owl>. In OWL DL a type for every individual is mandatory (beside 'owl:DeprecatedClass' or 'owl:DeprecatedProperty'). Any of you OWLites know what it's getting at in #2? Thanks, Tom Doman Novell Inc.
Attachment:
testLDAP.owl
Description: Binary data