From asun@delicias.dia.fi.upm.es Thu Oct  8 11:22 PDT 1998
From: asun@delicias.dia.fi.upm.es (Asuncion Gomez Perez)
Subject: Re: ontology tools?

...

Our tool is named ODE (Ontology Design Environment). It enables the developmen of 
ontologies at the knowledge level using the approach proposed by the methodology METHONTOLOGY.
The METHONTOLOGY framework and ODE allow the specification of ontologies at the knowledge level using a set of intermediate representations based on tabular and graph notations, and ODE translators transform and generate standard, consistent and well-strucutred ontologies in several target languages.  

METHONTOLOGY tabular and graph notations are a user-frinedly approach to 
knowledge acquisition and evaluation by domain experts that are not ontologists
and knowledge engineer. In particular, our experience shows that:

1. Domain experts and human final users do not understand formal ontologies codified
in ontology languages at all. For example, environmental experts do not understand 
ontologies in Ontolingua.

2. The same people, using the Ontology Server browser tools, could completely 
understand and validate taxonomies, partially understand instances, but could
not understand abstract definitions of concepts relations, fucntions and axioms.
Two environmental and chemical experts were able to understand a chemical ontology
and they validated the whole ontology in 6 hours.

3. They were not also to formalize their knowledge at all. 

4. However, such expert could understand, validate and formalize  80% of our intermediate represenations.

This is why we say that our approach allows to build ontologies at the knowledge level instead of at the symbolic level (as the ontology server does). It is not a matter
of changing the user interface (as MIke suggests).


ODE also includes a lot of checkings inside each and between intermediate representations. The goal is to prevent inconsistency and lost of knowledge during the 
conceptualization process. 

Another important difference between our tool and the Ontology Server, is that
we store the ontology in a relational database.  This has the big advantage that
future applications (integrated in the information system of a company) will 
access the ontology easily using SQL queries.

ODE is independent of any system capable of managing ontologies. That is, although
ODE generates Ontolingua code in ASCII, ODE does not interact with Ontolingua.

We have developed several ontologies using ODE:

.- A chemical-element and a chemical-crystal ontologies
.- We have  conceptualized the (KA)2 ontology using ODE and using ODE translators
we have generated the ontolingua code available at Madrid mirror site.
 Originally, the ontology was in Flogic. We spent just 8 hours in building the 
  ontology in ODE.
.- We built a REFERENCE-ONTOLOGY. It is a domain ontology about ontologies. 
.- At this moment we are building a monoatomic-ions ontologies that 
is going to be reused in a pollutant environmental ontology.
.- We also are building an ontology with a Knowledge Management application in mind.

ODE ontologies (in the relational DATabase format) are used by three applications:
.- Ontogeneration, a system that generate spanish texts using linguistic and domain ontologies.
.- (ONTO)2agent, an ontology-based www broker that helps to locate and select ontologies
.- A broker that uses the KM ontology.

If you are interested on some issues, you can find  the papers at 

http://delicias.dia.fi.upm.es/miembros/ASUN/asun_CV_Esp.html


Finally, ODe runs in a pentium with a Windows95 operating system. 

Best

Asun