[Sigia-l] RE: multi-faceted & topic map

[Lars Marius Garshol]

* Doug Howell
| 
| I recently asked the group what they thought of creating
| taxonomies/faceted classification systems from existing data models
| (entity relation diagrams), and got a few interesting replies. [...]
| Could you see creating an ontology from a data model, perhaps using
| the relationship names and types to define the relationship types in
| the ontology (or topic map)?

Structurally Entity-Relationship, especially as Chen originally
defined it, is *very* close to topic maps. In fact, it's basically a
subset of topic maps, and so you could easily convert an ER model into
a topic map ontology. With other ontology languages it is a little
harder, since they don't tend to model relationships the same way, but
generally it shouldn't be too hard.

You would probably have to tweak the model somewhat after doing this,
since traditional ER doesn't have class hierarchies, and ER models
also tend to be created for different purposes than ontologies, and
they also tend to relate to the real world in slightly different ways.

But in general this is something that's quite easy to do, and it also
makes a lot of sense if the ER model covers an area you'd like your
ontology to cover as well.

| How would this compare to creating taxonomies or FC systems from the
| same model?

Taxonomies and faceted classification are both much weaker than
Entity-Relationship, and they are also intended to do completely
different things, so any conversion here would have to be entirely
manual, and would usually involve a lot of thinking. I think in many
cases the ER model wouldn't really give you much help towards creating
such classification structures.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >