[protege-discussion] problem loading ontology

[Rui Lopes]

(Already posted to protege-owl, but still haven't got a reply. Apologies 
for the dupe).



Hi all,


I have encountered a strange error while loading an OWL ontology. This 
error occurs both by opening it from a local copy in my computer, or 
importing it from the Web into another ontology. The ontology in 
question is EARL (Evaluation and Report Language 1.0 Schema), as 
currently being standardised by the W3C.

The ontology location (and namespace) is http://www.w3.org/ns/earl#, 
whereas the current state of the spec can be found at 
http://www.w3.org/TR/EARL10-Schema/.

The error in question is as follows: (if required, I can post later the 
full exception stack as seen in the logs.)

"
There were errors at performing operation.

1. Exception class java.lang.RuntimeException.
Message: Illegal superclass DefaultRDFList(@_:A181 of 
[DefaultOWLNamedClass(rdf:List, FrameID(0:9109 0))]) of p1:Assertion has 
type class edu.stanford.smi.protegex.owl.model.impl.DefaultRDFList

2. Errors at loading OWL file from 
file:/Users/ruilopes/Documents/workspace/UUEval/ontologies/earl-temp.owl

Please consider running the file through an RDF or OWL validation 
service such as:
   - RDF Validator: http://www.w3.org/RDF/Validator
   - OWL Validator: http://phoebus.cs.man.ac.uk:9999/OWL/Validator


See console and log for more details.
"


I have successfully validated the ontology in the two suggested services 
(RDF and OWL validators). Moreover, I was able to use Jena 
programmatically to create an OWL ontology, append an OWL import 
construct pointing to http://www.w3.org/ns/earl#, serialise it to XML, 
query it through SPARQL successfully, etc. Therefore, it seems to me, 
imho, that it might be something weird happening in Protege.

However, the way the ontology has been specified is a bit unusual, since 
it uses OWL constructs only when there's no equivalent in rdf/rdfs 
(e.g., rdfs:Class, no owl:Ontology, etc.).


This bug/feature has been reproduced both in Protege stable (3.3.1) and 
beta (3.4), both behind proxy/firewall and open Internet environments. 
I'm using Protege in a Mac OS X 10.5.2 with 1Gb of RAM.

Is this a known bug of Protege? Is it a problem with the way the 
ontology has been specified (if so, it should be reported to W3C's 
respective WG)? Or is it something else incorrect from my side?


Please feel free to ask for more information, if required. Thanks in 
advance for the help!

Cheers,

Rui