Observations from the technical  paper presentations @Onto.com/ODISE 2014

1) Work on the FIBO ontology (Mike Bennett):

Interesting practical work in industry. In the end, he raised the issue of the need of a consistente methodology to evaluate ontologies.

The speaker mentioned the Oquare matriz – evaluation method that proposes some criteria for ontology evaluation.

We talked a bit about having a benchmark for ontology evaluation. Sergio de Cesare reminded us that evaluating an ontology is not a one-time thing. We must evaluate the ontology over a long period of time and using diferente datasets. If after 10 years, the ontology is still working on new and big-enough datasets, than it is stable, otherwise either it must be updated or the models should be updated.

2) Comments by Nicola Guarino at Maria Claudia Cavalcanti’s talk

-          Take the terms that are more common sense and assume them as primitives, focusing first on the more controversial terms

-          When working with the government, rely on the law (the used terms are usually founded on some normative act that should be considered when developing the ontology)

-          Usually, what the system technitians chose as terms is very distant than what is written in these laws.

-          Tiago Prince mentioned that even these laws are ambiguous by choice, not to close all doors (it is required for the law to be flexible) but even with these ambiguities, they help a lot.