Re: domain questionnaire

JRStern wrote:
> On 22 Feb 2001 17:20:42 GMT, hidders_at_REMOVE.THIS.win.tue.nl (Jan
> Hidders) wrote:
> >> > http://www.aw.com/product/0,2627,0201547325,00.html
>
> I just ordered both components from B&N.
>
> > http://www.acm.org/pubs/citations/journals/tods/1979-4-4/p397-codd/
>
> Errrr, 1979?

Yep, it's a classic. :-) But seriously, I actually feel that this paper was better thought through than Codd's paper on the original relational model. But because it wasn't picked up by the market the ideas were more or less ignored. And now they seem to be back.

> >No, ORM can be roughly described as a dialect of entity-relationship
> >models, but with a very good formal underpinning and an elaborate and
> >well-defined set of notations for constraints.
>
> ORM being Object Relational Model?

Er, sort of, it's Object Role Modelling. If you want to know more:

  www.orm.net
  www.inconcept.com

> I think it may be impossible to do
> this correctly and still see it as a dialect of relational -- and if
> E/R is something other than relational, I've missed it.

I'm not sure I understand this. Are you saying that there is no difference between the ER model and the relational model? And do you think that it is not possible to have an object-relational model?

> >> Have you ever read the philosophical works of Ruth Garrett
> >> Millikan?
> >
> >No, but the name sounds very familiar. Is her work similar to that
> >of Quine?
>
> Well, right now I am very, very high on Prof. Millikan. While there
> might be people who would say that they both advocate forms of
> philosophical naturalism, this is mostly a minimal commonality to
> twentieth century thinkers, a scientific worldview of one flavor or
> another. [...]

Well, your enthusiams has certainly wetted my appetite. I think I will move it up a few places on my 'must read sometime' stack.

> >Well, I wonder if that is really meaningful. How can I tell if
> >objects are different if I do not know any property or relationship
> >that it has with other objects?
>
> How can you compare two objects, if you don't have their identities
> separate before hand?

By having some way to refer to them. That is not the same as knowing how to identify them, i.e., determine if two references refer to the same object. You know, "morning star", et cetera.

> Classic relational theory puts the "identity" into the primary key,
> as an overload of their meaningful values. Why should this overload
> be a valid construction, or at least why should it be a necessary
> one, or at least, why shouldn't relational be mappable to an
> object-based structure in which these functions are kept separate?

Sure, you can, and as far as I know most formal object-based data models do. But the question is if you should introduce a design principle that says something like "for every class you have to have some predicate that is expressed in terms of the schema, that decides when two objects in this classe are one and the same object".

-- 
  Jan Hidders