Last night, the topic of dinner discussion was "ontologies." This subject is typically amorphous, so I thought I'd break it down into little manageable bites (which you'll see later is apropos).
In brief, Ontology refers to a structured information model. Usually this is described as a common vocabulary of terms or concepts and how they relate to each other. I'll come back to how this effects both computers and humans.
In philosophy, Ontology, according to Wikipedia, "is the study of being or existence and forms the basic subject matter of metaphysics. It seeks to describe or posit the basic categories and relationships of being or existence to define entities and types of entities within its framework. It is the science of what is, of the kinds and structures of the objects, properties and relations in every area of reality."
In information science, (getting back humans, like librarians and data specialists) ontologies describe the classification of data in entity relationships (ERs), or the conceptual representation of structured data. Okay, my professor in library grad school used to describe these relationships in terms of a information represented in a classification schema that is oft used in building taxonomies and thesauri. Doc Soergel referred to these schemas as the building blocks of information retrieval systems in database design. The basic (classical information science theory here) ER models are:
"Used For" (UF): Indicates that a term or concept is analogous, or a substitute.
"Narrower Term" (NT): describes a hierarchical relationship, obviously, a term or concept that describes a narrower than a previous concept. The concept of "fruit" is narrower than "tree" which is narrower than "botany" which is narrower than "science", for example.
"Broader Term" (BT): the inverse of a Narrower Term.
"Related Term" (RT): this is purposefully vague--usually refers to a term or concept that is not hierarchically related to another term. Not the same as UF.
Typically, this view of information is very structured...a lot of the cool kids don't like to use it these days, but as a librarian-person, I appreciate the simplicity and the fact that it is
still used.
What do the cool kids think about ontologies and information relationships? Well some still still use the classical model to build databases (heh, you have to retrieve information some how), but others describe
non-classification approaches like this oft-cited guy. To reduce the long-rambling, but interesting article, Mr. Shirky feels that there is no ideal classification schema, and no way people can agree on one. Shirky thinks that we are moving away from a binary information approach to one that's more organic, like tagging or
folksonomies.
So what to do? Well you can argue that if you put Soergelettes and Shirkyiods in a room together, they'll either destroy each other or come up with the perfect schema. Or maybe they'll agree to coexist as RTs. I'll get back to you.
