CEL: OWL 2 Implementation report -- update

Hello,


    1.  Your name, affiliation, and (optionally) the names of other
        people who helped with the implementation.

Boontawee Suntisrivaraporn, SIIT, Thammasat University, Thailand
Anni-Yasmin Turhan, TU Dresden, Germany
Julian Mendez, TU Dresden, Germany


    2.  The name of your system, a URL for its website (if any), and a
        one-sentence description.

CEL (see http://code.google.com/p/cel/ )

CEL implements, unlike other modern DL reasoners, a polynomial-time
algorithm for the classification of EL ontologies.




    3.  Which profile(s) it implements (DL, EL, QL, RL, or Full).  We
        would appreciate some brief commentary about why you chose those
        profiles, and what sort of implementation techniques you are
        using.

CEL implements (a subset of) OWL EL only.

CEL implements a polynomial time algorithm (see [1]) specifically 
tailored to EL (or rather EL++) the actual implementation is described 
in [2]. CEL also supports supplemental reasoning features like 
incremental classification, modularization and axiom pinpointing. 
Moreover, the OWL API wrapper for CEL has eventually become available, 
so now you can use CEL as the backend reasoner from within Protege.


[1] Franz Baader, Sebastian Brandt, and Carsten Lutz.  Pushing the EL
Envelope Further. In Kendall Clark and Peter F. Patel-Schneider,
editors,  In Proceedings of the OWLED 2008 DC Workshop on OWL:
Experiences and Directions, 2008.

[2] Franz Baader, Carsten Lutz, and Boontawee Suntisrivaraporn. Is
Tractable Reasoning in Extensions of the Description Logic EL Useful
in Practice?. In Journal of Logic, Language and Information, Special
Issue on Method for Modality (M4M), 2007.


    4.  Which semantics you implement (direct or rdf-based), and
        (optionally) why.

CEL implements direct semantics.


    5.  Do you believe your system currently conforms to the OWL 2
        Candidate Recommendation?  Does it pass all the test cases for
        your profile?  If not, which features does it lack and/or which
        test cases does it not yet pass?  Do you have plans to make it
        conformant, and make it pass all the test cases?

CEL is conformant with those parts that are implemented in the system.
CEL cannot handle ontologies that contain nominals and safe concrete 
domains (thus, one-of, datatype properties and feature chains in OWL).

CEL passes all tests that do not include the above mentioned unsupported
feature. For the next major version of CEL it is planned to include the
missing features.


    6.  Did you implement the "at risk" features, owl:rational and
        rdf:XMLLiteral?  If not, do you intend to, or do you think we
        should remove them from OWL 2?

CEL does not implement these features.




    7.  Finally, we'd appreciate your evaluation of whether the OWL 2
        Candidate Recommendation is ready to proceed along the standards
        track toward being a W3C Recommendation.  If not, please be sure
        to tell us what problems you think we need to address.

We believe OWL2 is ready to proceed to Recommendation.


Regards, Anni


-- 
*  Anni-Yasmin Turhan, Theoretical Computer Science, TU Dresden    *
*  Phone:++49 351 463 39 167  mailto:turhan@tcs.inf.tu-dresden.de  *

Received on Thursday, 6 August 2009 09:05:54 UTC