[seek-kr-sms] Loading ontologies without internet connectivity

Shawn Bowers bowers at sdsc.edu
Wed Nov 24 07:38:46 PST 2004


On Tue, 23 Nov 2004, Rich Williams wrote:

> Hey Shawn -
>
> Not sure what you mean by a solution in terms of the organization of the
> ontologies.  The only one I can think of would be to have everything in one
> file, and I kinda like having it in multiple files, helps me keep track of
> where things are, and perhaps one day will be useful to someone who for
> example wants to use our model of measurements (and units etc) for something
> non-ecological.

My question was if it were possible to divide the ontology represented by
the current files into different files that don't require as much
importing; i.e., a more "stand alone" file organization. Alternatively,
perhaps a more fine-grained file organization would be useful, in which
only one "top-level" file is imported into each smaller, thematic file.
The thematic files would contain only a handful of concepts.

I wouldn't necessarily argue for a single file approach, but that is
pretty much what using imports does ... e.g., to use measurement base, all
the imports must be loaded as well, and the imported file's imports, and
so on.

I think the bigger question is how best to package the ontology
represented by all of these files to a user. Should the file organization
be meaningful, or should there be some other mechanism to easily present
useful portions of the ontology?

With database schemas as well as ontologies, I think the 7 +- 2 rule
generally holds: it is difficult for one to process more than 5 - 9
concepts (or properties, etc.) at one time.

For the ont-policy, I would have assumed you had this stored within your
cvs directory, and that paths were relative ...  but perhaps the relative
url's don't work.

Thanks Rich,

Shawn


>
> As for ont-policy, I use it for both my work with Protege and for the
> applications I'm developing that use Jena.  It doesn't really seem like
> something I should check in - it is specific to my hard disk structure.  I
> just added a bunch of declarations like this:
>
> <OntologySpec>
>     <!-- local version of the FoodWeb ontology Properties.owl -->
>     <publicURI
> rdf:resource="http://wow.sfsu.edu/ontology/rich/Properties.owl" />
>     <altURL
> rdf:resource="file:///D:/nceas/seek/projects/kr-sms/OWLOntologies/Properties
> .owl" />
>     <language  rdf:resource="http://www.w3.org/2002/07/owl" />
>     <prefix    rdf:datatype="&xsd;string">prop</prefix>
> </OntologySpec>
>
> to the default ont-policy file, one for each owl file that I keep a local
> copy of.
>
> Rich
>
> > -----Original Message-----
> > From: Shawn Bowers [mailto:bowers at sdsc.edu]
> > Sent: Tuesday, November 23, 2004 4:21 PM
> > To: Rich Williams
> > Cc: seek-kr-sms at ecoinformatics.org
> > Subject: Re: [seek-kr-sms] Loading ontologies without internet
> > connectivity
> >
> >
> >
> > Rich,
> >
> > Thanks for the response. I am actually doing this now in Jena, but I
> > wonder if there is a more general solution in terms of organization of
> > the existing ontologies?
> >
> > Also, do you use ont-policy.rdf for protege, and if so, can you check it
> > into cvs?
> >
> > thanks,
> > shawn
> >
> >
> > Rich Williams wrote:
> >
> > > This is an issue in GrOWL, but it's really an issue with the
> > OWLAPI.  There
> > > needs to be a way to tell it an alternative location of a
> > resource when the
> > > uri is not available.  You can do that for the main file (by
> > loading it from
> > > your local disk) but not for the imports.  Such a mechanism
> > exists in Jena
> > > (the  ont-policy.rdf file), so you can work with a local copy
> > of an ontology
> > > and its imports without changing resource names (I do it all
> > the time when
> > > editing in Protege).
> > >
> > > Rich
> > >
> > >
> > >>-----Original Message-----
> > >>From: seek-kr-sms-admin at ecoinformatics.org
> > >>[mailto:seek-kr-sms-admin at ecoinformatics.org]On Behalf Of Shawn Bowers
> > >>Sent: Tuesday, November 23, 2004 8:52 AM
> > >>To: seek-kr-sms at ecoinformatics.org
> > >>Subject: [seek-kr-sms] Loading ontologies without internet connectivity
> > >>
> > >>
> > >>
> > >>I am having some problems with GrOWL and Jena with Rich's Ontologies
> > >>when there is no internet connectivity.  And I was wondering if anyone
> > >>had run into this, or had ideas on how to fix this ...
> > >>
> > >>When I am not connected to the internet, and I try to load an ontology
> > >>such as Ecological Concepts, GrOWL dies.  Jena also has problems. In
> > >>particular, all of the concepts such as EcoOntThing that are from an
> > >>imported ontology become problematic.  For example, exceptions are
> > >>thrown if these are treated as OntClass instances or if imported
> > >>properties are treated as OntProperty instances.
> > >>
> > >>One solution to this problem might be to import local files instead of
> > >>from URLs.  Does anyone have any other solutions?
> > >>
> > >>In general, I think we might want to have a more self-contained and
> > >>possibly thematic organization to the SEEK ontologies.  Rich, have you
> > >>thought about how best to organize what you've done? Or, is the current
> > >>organization your preferred approach?
> > >>
> > >>Shawn
> > >>
> > >>
> > >>_______________________________________________
> > >>seek-kr-sms mailing list
> > >>seek-kr-sms at ecoinformatics.org
> > >>http://www.ecoinformatics.org/mailman/listinfo/seek-kr-sms
> > >
> > >
> > > _______________________________________________
> > > seek-kr-sms mailing list
> > > seek-kr-sms at ecoinformatics.org
> > > http://www.ecoinformatics.org/mailman/listinfo/seek-kr-sms
>
> _______________________________________________
> seek-kr-sms mailing list
> seek-kr-sms at ecoinformatics.org
> http://www.ecoinformatics.org/mailman/listinfo/seek-kr-sms
>




More information about the Seek-kr-sms mailing list