Re: SWRL Implementations (Concerning Pellet & SWRL-builtins)

Hi,

Apparently pellet is almost the only reasoner that have formal and sound
implementation of SWRL and not a transformation to another rule language.
However it misses reasoning with SWRL builtins so far.

Until reasoning builtins is added to Pellet,  can one define the needed
built-ins as general rules in the application ontology??

Regards,
Somaya

On Nov 20, 2007 1:23 PM, Somaya Aboulwafa <somaya.ahmad@gmail.com> wrote:

> Many thanks for all your replies.
>
> I have two questions here, you said that: "The SWRLTab performs inference
> using a Jess back end and while the implementation supports nearly all of
> SWRL it does not consider all OWL axioms during inference. However, the
> latest Protege-OWL 3.4 beta also has a direct connection to the Pellet
> reasoner, which has a much more formally sound implementation of SWRL, and
> is also under active development. However, Pellet has limited built-in
> support."
>
> Does this mean that if I used SWRLTab with Jess back-end then I won't be
> able to reason over all the OWL axioms in my swrl ontology? If so, is there
> any workaround to this problem?
>
> Also if I used SWRLTab with Pellet then I won't be able to make use of the
> SWRL-builtins implmemntation made by SWRLTab?!
>
> Thanks and regards,
> Somaya
>
>
>   On Nov 19, 2007 4:25 AM, Martin O'Connor <martin.oconnor@stanford.edu>
> wrote:
>
> >
> > The SWRLTab in the current Protege-OWL 3.4 beta supports most of the
> > core SWRL built-ins [1] and has an expanding set of other built-in
> > libraries [2]. Protege 3.4 has a SWRL/RDF parser; Protege 4 alpha also
> > has a SWRL/OWLX parser, though the SWRLTab will not be available in
> > Protege 4 for several months. The SWRLTab also has a SWRL-based OWL
> > query language called SQWRL [3]. All SWRLTab functionality is
> > accessible through a fully documented set of Java APIs [4].
> >
> > The SWRLTab performs inference using a Jess back end and while the
> > implementation supports nearly all of SWRL it does not consider all
> > OWL axioms during inference. However, the latest Protege-OWL 3.4 beta
> > also has a direct connection to the Pellet reasoner, which has a much
> > more formally sound implementation of SWRL, and is also under active
> > development. However, Pellet has limited built-in support.
> >
> > I'm superficially familiar with the other systems you mention but I do
> > not know them in enough detail to answer your questions. Perhaps
> > others here can.
> >
> > Martin O'Connor
> >
> > [1] http://protege.cim3.net/cgi-bin/wiki.pl?CoreSWRLBuiltIns
> > [2] http://protege.cim3.net/cgi-bin/wiki.pl?SWRLTabBuiltInLibraries
> > [3] http://protege.cim3.net/cgi-bin/wiki.pl?SQWRL
> > [4] http://protege.cim3.net/cgi-bin/wiki.pl?SWRLTab
> >
> > Quoting Somaya Aboulwafa <somaya.ahmad@gmail.com>:
> >
> > > Hi,
> > >
> > >  I have recognized some swrl implementations like Pellet, Bossam,
> > KOAN2,
> > > RacerPro, SweetRules (Jena2 rules), SWRLJessTAB (in protege),
> > R2ML(REWERSE).
> > > And i need to choose one that have the following features:
> > >
> > > 1- Applies most features of SWRL, like (SWRL/OWLX Parser, SWRL/RDF
> > Parser,
> > > SWRL Math Built-Ins, SWRL String Built-Ins, SWRL Comparison Built-Ins,
> > SWRL
> > > Boolean Built-Ins, SWRL Date, Time and Duration Built-Ins, etc.). I
> > know
> > > that almost all of the above implemenattion takes only a decidable
> > frgment
> > > of SWRL. but i need the one that takes the largest fragment without
> > going
> > > through the decidability problem.
> > > 2- Reliable resoning; no need for manual checking of the results.
> > > 3- Java API where i can use it from within java applications.
> > >
> > > Thanks and regards,
> > > Somaya
> > >
> >
> >
> >
>