[sem-grd] CIM

Bartosz Kryza bkryza at icsr.agh.edu.pl
Wed Oct 19 09:51:38 CDT 2005


Hi,

I just wanted to share some of my expierences with converting CIM to OWL.

Within the framework of K-Wf Grid project, which I am a member of, I've 
developed a small tool that was supposed to convert CIM schema to OWL.
I based it on the reasearch from the paper mentioned by Dave and on 
grammar for MOF (Managed Object Format) in which CIM is defined, which 
is available from Pedro Assis and can be found at 
http://www.antlr.org/grammar/cim.

Thanks to this grammar, implementation of tool itself is not the 
problem, but the real problem is definition of a good mapping from CIM 
to OWL, which as mentioned in this paper cannot be perfect.

I've deployed some time ago the OWL files generated by my tool, however 
please keep in mind they are still far from being usable:
http://gom.kwfgrid.net/cim

Some of them, e.g. Core or Application can be loaded with some latest 
version of Protege... however some of them might fail. Also, please 
forgive any 'dumb' mistakes inside these ontologies, these are caused by 
the fact that they are generated automatically, and they're size makes 
validation quite difficult.

Anyway, I agree that converting at least a subset of CIM to OWL is of 
crucial importance if we want to think of reasoning using concepts 
defined in CIM schema.
Of course I'm planning to finalise the tool as soon as I define 
satisfactory mappings.

Hope You find this useful.

Best regards,
Bartek Kryza
ACC Cyfronet-AGH

David De Roure wrote:

>Some of you will be aware of the interest in representing all or part of the Common Information Model (CIM) in RDF(S)/OWL - this is a potentially useful step which also provides a real experiment in applying Semantic Web to Grid.
>
>The following paper may be of interest
>
>'Toward a Formal Common Information Model Ontology'
>
>w3.antd.nist.gov/pubs/quirolgico-wise2004.pdf
>
>I've also been looking at tools which can represent UML in W3C formats.
>
>I'd be interested to explore this further and welcome contributions from others.
>
>Thanks
> 
>
>-- Dave
>
>
>
>  
>





More information about the sem-grd mailing list