Re: Roles and types (yet again)

* Lars Heuer
|
| Hi Steve,
| 
| > | Is Joe Cocker an instance of "plumber", "composer", or "musician"? In
| > | a better model these instance of statements would be modelled as
| > | associations where the mentioned topics would i.e. play the role
| > | "occupation".
| 
| > And just what would the associations be in this case, as a matter of
| > interest?
| 
| I.e.
|     has-occupation(person: joe-cocker, occupation: plumper)
|     has-occupation(person: joe-cocker, occupation: composer)
|     has-occupation(person: joe-cocker, day-job: musician)

OK. I could accept that. Other proposals I've seen would have me specify one or more
"composed by" associations in order to express that someone was a "composer". I
don't accept that.

But even though your solution is acceptable, I think it's wrong to always insist on
drawing a sharp distinction between role type and topic type. The categories of
human knowledge are better expressed using a prototype model than a
criterial-attribute (or "Aristotelian") model. In such a model, roles and types are
not sharply differentiated, but rather exist on a continuum. Thus, to categorize Joe
Cocker as a plumber makes less sense than categorizing him as a singer; categorizing
him only as a person is not very interesting at all.

Steve