new release of ontology library
Tom Gruber <gruber@HPP.Stanford.EDU>
Message-id: <199408040551.WAA10363@HPP.Stanford.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Wed, 3 Aug 1994 22:57:29 -0800
To: srkb@cs.umbc.edu, ontolingua@HPP.Stanford.EDU, interlingua@isi.edu
From: Tom Gruber <gruber@HPP.Stanford.EDU>
Subject: new release of ontology library
Sender: owner-srkb@cs.umbc.edu
Precedence: bulk
There is now a fresh set of ontologies in the Knowledge Sharing Effort
library at http://ksl-web.stanford.edu/knowledge-sharing/ontologies
What's new:
- several new entries: ontologies for component assemblies with
connections and parts, parametric constraints, extensions into geometry and
tensor algebrae
- new and improved domain theories: VT (elevator configuration using the
components and constraints ontologies - NEW VERSION), MACE (parameteric
satellite model), and thermodynamic systens (translated from the CML model
fragment library).
- better documentation. There are now publications that cite the online
ontologies (and hypertext links between natural language documents and
formal definitions of terms in ontologies).
- the ontologies live in hierarchical theory namespaces, so that there is
reuse of ontologies within the library (axcioms inherited from parent
theories) while allowing for multiple namespaces (disjoint theories in the
hierarchy). Although this is not a context mechanism (which we would hope
would come form the KIF group), it is a practical way for us to share
ontologies without having them all designed and maintained together.
- all the released ontologies check out under a variety of tests, and all
are load without error into all of the target representations. Notable
achievements are Loom (which is expressive enough to be dangerous) and
Clips (which is limited but is widely used and is representative of OO
systems).
The library is open to all to contribute. We have received some other
ontologies, but they are still full of errors. The new ontolingua release
should facilitate in the generation of portable ontologies. We will be
glad to post any ontology that
(1) loads into ontolingua and passes the XREF analysis without warning
(2) comes with a text file introducing it (we will automatically link the
text file to the formal definitions in the ontologies).
The preferred place for discussions about ontologies themselves is
srkb@cs.umbc.edu
and the mailing list for ontolingua-specific issues is
ontolingua@ksl.stanford.edu
Archives of these lists, as well as the related lists interlingua@isi.edu
and kqml@cs.umbc.edu, are available on the Web at
http://ksl-web.stanford.edu/email-archives/
tom gruber