[seek-dev] [Bug 1258] New: - store, query, and access ontologies in ecogrid
thau at learningsite.com
Thu Jan 15 14:39:03 PST 2004
Just to throw in my own OWL tool experiences - I'm finding that Protege
has the same problems as Rich describes below. It changes cardinalities
around, (e.g. all of a sudden my taxa can have more than one rank) and it
changes the types of properties too (e.g. suddenly, has_child and
has_parent are no longer inverses).
So, now I'm doing first drafts in Protege and editing them by hand and
then looking at them in Protege to see if they still look right.
In terms of reasoning engines and consistency checkers, I haven't gotten
FaCT to work on my Linux box, and Racer still has the problem that it
assumes that all individuals are different, which violates a major OWL
tenet. I'm currently using Pellet, which is from mindswap (like some of
the tools Rich describes) to check the consistency of my knowledge bases
and doing reasoning over them. I'm pretty happy with it. It does a good
job of telling me why my ontology doesn't conform to DL rules, it's
pretty quick, and it seems to score very well here:
On Thu, 15 Jan 2004, Bertram Ludaescher wrote:
> Thanks for sending this nice summary. I'm cc-ing Shawn and Kai (from
> GEON) explicitely too...
> >>>>> "RW" == Rich Williams <rwilliams at nceas.ucsb.edu> writes:
> RW> Hi -
> RW> I saw this bug posting in my email and thought I could provide some helpful
> RW> context. There are some tools available in the semantic web community that
> RW> address many of the needs of the ecogrid system, though of course not in a
> RW> grid computing environment.
> RW> Jena (http://jena.sourceforge.net/) is a set of open source Java tools that
> RW> grew out of the HP semantic web effort. They were chosen by the Protege
> RW> group when they decided to add OWL support to Protege, a choice that I
> RW> believe is indicative of the relative maturity of the Jena tools. Jena
> RW> includes parsing of RDF and OWL from various text representations,
> RW> persistent (database) storage of RDF triples, RDF and OWL APIs, an RDF query
> RW> language, and support for inference engines. I recently began working with
> RW> the Jena APIs and found them relatively easy to learn and it has so far been
> RW> reliable.
> RW> There are various European based semantic web initiatives that provide tools
> RW> in various states of development. Check out
> RW> http://wonderweb.semanticweb.org/ and http://www.ontoknowledge.org/. I've
> RW> been adapting the WonderWeb OWLAPI for use in the OWL version of
> RW> OntoBrowser. It definitely shows its 'alpha' status - the code for reading
> RW> ontologies and exposing them in a set of Java classes is well developed, but
> RW> the code for writing the ontology back out in OWL/RDF is not as reliable and
> RW> I've had to fix some bugs in it.
> RW> I hope this is useful.
> RW> Rich
> RW> _______________________________________________
> RW> seek-dev mailing list
> RW> seek-dev at ecoinformatics.org
> RW> http://www.ecoinformatics.org/mailman/listinfo/seek-dev
> seek-dev mailing list
> seek-dev at ecoinformatics.org
More information about the Seek-dev