[Sigia-l] Question about the use of the term 'ontology'

[Lars Marius Garshol]

* ruth at ruthkaufman.com
|
| Background:
| I'm writing a project definition document for a project concerning
| the development of standards for product descriptions. A full set of
| standards may include editorial guidelines, presentation standards,
| interaction patterns, an XML schema or DTD, policies, controlled
| vocabularies -- anything that everyone in the organization must
| enable or comply with in order to make product descriptions sharable
| among systems and rendered consistently for users. I anticipate that
| we'll need to create or modify some controlled vocabularies.  These
| may be corporate taxonomies, allowed values lists, or some other
| kind of -something-.

That's interesting. It doesn't really sound like "standard IA" to me.
What's the context of this? Some kind of product data management, like
product configuration, or more general?
 
| Question:
| Is 'ontology' the overarching term here for 'words for kinds of
| things'? 

Not really, but I do agree with Peter here: usage of this term is all
over the place.  There's one meaning of this in philosophy, another in
library science, and a third in computer science. The latter two
overlap to some extent, and I would describe the greatest common
denominator as something like:

  "an ontology is a model of a subject domain consisting of entity
  types, property types, and relationship types, where the entity
  types usually have some degree of class hierarchy, and the model is
  meant to reflect the domain more or less directly as it is"

| I basically need one generic, all encompassing term to cover strict
| taxonomies and other kinds of controlled vocabularies and lists of
| things. 

That wouldn't be "ontology". In library science that would be
"subject-based classification". Outside of library science I don't
really know of anything that works.

| In this post, I've used 'controlled vocabularies', but I have a
| vague recollection that this refers to something specific and isn't
| the best overarching term (or is it?). 

You could stretch it to fit, but it's better not to, I think.

| I have another notion that controlled vocabularies and taxonomies
| simply reify ontologies, in which case ontology wouldn't really be
| the right category heading... unless there are two meanings for
| 'ontology' -- one philosophical and one used by IT & knowledge
| workers.

I would go with "subject-based classification", but, like Peter
suggests, take care to define it.


I actually wrote a paper which attempts to clear up the terminological
mess in this whole area (metadata, classification, ontologies, ...),
which you may find interesting. It's at:

  <URL: http://www.ontopia.net/topicmaps/materials/tm-vs-thesauri.html >

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