No subject

programming (RD languages such as SQL, , Relational modeling languages
such as ERD).

OODBMS is a data storage system and an inference engine (query engine)
built around the paradigm named Object Oriented (centric). From this
paradigm several languages exist both for modeling and programming (OOP
languages such as Java and OQL, OOM languages such as UML and ORM). 

>Ontologies is just a generic name for
knowledge-base-that-models-knowledge.

Both of these paradigms (OO and ER) were built to model knowledge too (a
paradigm can be seen as the knowledge base for modeling knowledge). I
guess I'm looking for the paradigm sustaining what the "Ontologists" are
inventing. Or can we say that both OO and ER are in fact paradigms that
can be used also to define Ontologisms? If not, what are their flaws? In
other words what kinds of artifacts are missing in their paradigms?

>The actual structure of an ontology can differ (there is no ONE
structure >for ontologies. XTM (topicmaps) provides one example of a
defined structure >that can be used to model ontologies. Topicmaps are
often kept in a >relational database rather than in the XML format
(XTM).

I see. Can you provide me with links to "open source" projects that we
can see how the concept of generic topicmaps was mapped into an
entity/relational model (if they are stored in a relational database
then the proof of equivalence is made at least in part (derivation may
be)).

But again it seams there is a paradigm behind the definition of
Ontologisms that is different from OO and ER. Usually from an
engineering point of view a paradigm is defined with a set of artifacts
to model knowledge (information and its relationships) plus another set
of artifacts that provide a systemic approach to reasoning around the
model (systems of proof).

>A specific Ontology structure can be modelled with objects.

I see. So, not all Ontologisms can be modeled with objects using the OO
paradigm. Can you provide an example of such and that is used regularly?

>For a simple hierarchical faceted system, try:
>Facet table
>- id
>- name
>Topic table
>- id
>- name
>- parenttopicid
>- facetid

I see. You are modeling data using the ER paradigm. But how would the
inference engine work?

>How 'bout http://xfml.org

Yes, I was thinking about that also.

>Like the name :)

The name was just a coincidence, or probably not. After all we all live
in the same world :)


I'm starting to suspect that there is no general paradigm around the
definition of Ontologies, an IA is not delivering it also. But that is
not bad at all (it's even excellent otherwise I would be stuck on my
attempt to develop a generic system using a mix of ER and OO). It seams
to me that Topic Maps, Faceted Classification and other things are
essentially structures and algorithms (strategies) used to map
Ontologies to an artificial system. But I'll keep on with my eyes open.

>Hope that helped.

Yes indeed. Thanks a lot Peter for your time.

Best regards,

Nuno Lopes