Just create a new OWL Project from existing sources and
paste the above URLs.
Every package was transformed to a sub-ontology (1 Package
– 1 File). Dependencies and inter-references between packages represent
the imports. The navigableassociation-ends, as well as the
attributes where mapped to object properties – with respect to their
range, domain and cardinality. We didn’t use the xsd datatypes for
representing the common datatypes (String, Integer, UnlimitedNatural and
Boolean) – they were represented as classes. The Enumeration classes were
mapped to owl:Class; their attributes (possible values) were transformed to
instances of the class and explicitly made disjoint (owl:AllDifferents).
One limitation is the representation of the behavioral features
(operations and parameters); they were converted to rdfs:Comment. All the
operations have a private visibility and their specification is an OCL
Expression – but expressed in Text, not XML constructs.
The aim of this transformation is to precisely describe the
OMG specs and store systems specification expressed in UML or MOF in an OWL
database, instead if using a mof-based repository.
All comments will be appreciated.
P.S: The recommended version of Protégé is 3.2Beta build
304. Build 318 crashes when selecting the active sub-ontology (and there are 82
files for UML2).