Re: Principle of Orthogonal Design

On Jan 22, 9:36 pm, "David Cressey" <cresse..._at_verizon.net> wrote:
> "JOG" <j..._at_cs.nott.ac.uk> wrote in message
>
> news:87fece57-0892-4a5f-b3c0-a06c03ccc85b_at_e4g2000hsg.googlegroups.com...
>
> > On Jan 22, 7:07 pm, "David Cressey" <cresse..._at_verizon.net> wrote:
> > > But "home phone number // work phone number" sure looks like gen-spec
> to
> > > me.
> > > "phone number" is data that can be applied to a "telephone"
> generically.
> > > a "home phone" is a specialized phone. (Actually, it isn't the
> telephone,
> > > but the telephone line or more specifically the associated service
> contract,
> > > that is peculiar to either home or work).
>
> > Keep your dastardly OO away from my query neutral data, and leave it
> > in Smalltalk where it belongs ;)
>
> I plead not guilty. ER modeling sometimes has the gen-spec pattern added to
> the other patterns.

If anything is going to tie one into a single conceptual model, and generate application bias a-go-go, it is going to be arbitrary is_a hierarchies. ooooh, they make me mad.

>
> See the U Texas data modeling website.
>
> > Given:
> > - Frank's home telephone number is X
> > - Franks work telephone number is Y
> > - Tom's home telphone number is Z
>
> We don't need "numbers" in the context of this discussion.
>
> What's not clear to me is whether we have any attributes that are unique to
> home phones or to work phones.

But why would you even /want/ to try and arbitrarily chop up the world, inventing object types with artificial boundaries and convoluted inheritance trees. Why not just state that "To call Frank at Home we dial the number X" and leave it at that? Why make it any more complicated?

I am starting to conclude that we have evolved an absolute obsession to categorize and partition, to chop up the world into little pieces, to the point of ocd. I certainly suffer from this affliction, and I imagine it is almost self-fulfilling prerequisite for anyone who ends up getting involved in database theory.

But deep down, I really blame soduko. Damn them number squares.

>
> That depends on the U of D. Also, "Frank's work telephone number is Y"
> might really mean "Frank can be reached, at work, at telephone number Y".
>
> Oops, I forgot. We aren't supposed to be discussing meaning.
Received on Wed Jan 23 2008 - 00:23:21 CET